applied ML research and projects

Technical Directions

Machine learning and some form of pattern recognition or statistical analysis and classification is threaded throughout my career and free endeavors. In particular, fusion of context (salience or a general pattern through digital content (audio, video, or image) or loose textual description) and behavioral prediction (for understanding of history and guidance to what's next) are present in many forms of new technology and entertainment. I have participated in research endeavors in both multimedia and natural language processing and am always excited to discover and implement new approaches to non-traditional problems.


Detecting and generlizing behavioral experiences from users through explicit and passive observations.


Given text, aural, visual, and gestural inputs, frame general concepts and specific objects of interest.


Experiences are ephemeral but preferences may persist.  Find the best combination of both while  utilizing the trends of cohorts where necessary.


Everyone is a developer and creator, but needs assessments and feedback are critical to finding the ideal solution for everyone invovled.


Projects mark milestones for significant accomplishments and offer checkpoints for user interactions and business use cases.  Select projects are included below.


Call Before You Dig (2021-2023)

With a query rate of 19 million tickets a year for the location buried cables and facilities, a fast and accurate response is critical.  In a multi-year staged improvement process, combined data science and field operations teams sought to bring automation novel modeling and semi-automated rule creation. This research combines geospatial recognition, natural language processing, and computer vision to help maintain uninterrupted customer service while adhering to customer friendly response standards and exacting regulatory timeline requirements.  Upon completion of the project, a persistent saving of 7% YoY was achieved with systematic training data updates and in-field validation.


snackable video and AI-based edits (2021)

To accelerate the creation AI-generated "snackable" video clips, we explored the automated curation of "snackable" content from movies and TV shows.  This peer-reviewed user-preference study evaluated both human and machine learning-based curation methods. The study reveals a user preference for slightly longer curated clips that feature human characters either at the beginning or end, challenging initial expectations about ideal snackable content durations and composition. This work with Allyson King also offers insights on preferred automated methods for producing these short content pieces in alignment with user tastes.


think wisely (2020)

Wisely (aka ThinkWisely) was a lanaguage model that offered to be a friendly mediator.  In the early 2020's, the rise of vehement opposition and disbelief of media was a detriment to civil discource and discussion.  Trained from an interactive game that solicits contrarian interactions, this sytem would use a number of semantic labels (e.g. bullying, aggressive, unfounded) to detect and warn users when a source's conversation may go afoul of the recipients expectations.  Working with James Pratt and Connie Missimer, this project applied for an SBIR-Phase1 grant, but no contributor could step aside as a founder, so the launch point passed.


BotsAlive (2018)

BotsAlive was created as a way to inspire children to embrace technology and engineering.  Playful immersion in advanced STEM. Using Hexbug Spiders, teachable robots solve kid-built mazes with artificial intelligence. Challenges involved using a computer vision solution powered only by a local mobile device and using reinforcement learning to teach and evaluate optimal paths several times a second.  Working with Brad Knox, we applied for and won an SBIR-Phase-1 grant (as Emoters) focusing on a program that would dynamically learn and inspire from interactions with robots.


CuZero (2008)

CuZero is the culmination of prior interactive search work and the centerpoint for my thesis, aimed at two problems: difficulty in formulating a query and difficulty in browsing many queries at a time. Joint human-computer queries guides a user to the best concepts or query parameters for their search. A new image browsing strategy simultaneously allows a user to view many different query permutations with a visual grid for fast re-weighting and inspection. Following in the steps of CuVid, CuZero logs user interactions to further automate the process of search and inspection of new images in a large database.

Visual Islands illustration and wofkrlow

Visual Islands (2008)

Visual Islands change the way we inspect results and was graciously given the best oral paper award at CIVR 2008. With Cheng-Chih Yang, we asserted  incorporation of semantic or other meta-data, would offer the user an alternative layout for images that would allow faster comprehension and identification of related images. We also proposed a method that allowed non-linear browsing of results by finding relevance images and hopping to that page of results directly.

CuVid Semantics

CuVid (Sematics) (2006)

A joint work with Lyndon Kennedy involved the full exposure to our concept library in different ways. We incorporated several semantic search options (i.e. lexical similarity, etc.), options to find related items by dominant concepts (published as a joint work with AT&T above), and options to recommend filters using these dominant concepts.   This filtering can be performed individually, or together as a single process.


CuVid (feedback) (2006)

CuVid first incorporated multi-modal search in what has become a standard fashion. User interaction was used to dynamically rerank results with our published CueX reranking algorithm.


Historical publications are included with category tags and synopsis documents prepared by a GenAI analysis of the abstract and introduction.

Copyright Notice: This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.

2023-07-13 - A General-Purpose Multi-Modal OOD Detection Framework; Viet Duong, Qiong Wu; Zhengyi Zhou; Eric Zavesky; Jiahe Chen, Xiangzhou Liu, Wen-Ling Hsu, Han Zhao, Huajie Shao; pendingOut-of-distribution detection, multi-modal, machine learning, binary classifier, contrastive learning, Hinge loss, anomaly detection, feature sparsity regularizer, scoring metric, general-purpose frameworkThe proposed WOOD framework leverages a combination of a binary classifier and a contrastive learning component, using Hinge loss to better discern between in-distribution and out-of-distribution samples. A new scoring metric is introduced to fuse predictions from both components, enhancing the detection of multi-modal OOD samples under various real-world scenarios. This framework outperforms existing methods in simultaneously detecting different types of OOD data across multiple datasets, demonstrating its generalizability and applicability.
2021-04-21 - User Preferences for Automated Curation of Snackable Content; Allyson King; Eric Zavesky; Michael Gonzales; IUI '21: 26th International Conference on Intelligent User Interfaces, April 2021, Pages 270–274 [PDF]snackable content, automated video curation, machine learning, user preferences, human curation, content personalization, social media, entertainment content, video cutting, empirical studyThis paper investigates user preferences for the automated curation of "snackable" content from movies and TV shows, evaluating both human and machine learning-based curation methods through comprehensive user studies. The study reveals a user preference for slightly longer curated clips that feature human characters either at the beginning or end, challenging initial expectations about ideal snackable content durations and composition. The paper also offers insights on preferred automated methods for producing these short content pieces in alignment with user tastes.
2021-04-07 - Curating Short-Form Content from Views and Activity Classifiers; Jianxiong Dong; Subhabrata Majumdar; Tan Xu; Mohammed Abdel-wahab; Eric Zavesky; Data Science Practicum
Dr. Shannon Quinn, University of Georgia
Snackable Media, Video Action Recognition, Viewership Analysis, Content Summarization, Personalized Curation, Short-form Content, Automatic Recommendation, Audio-Video Content, Content Consumption, Data ScienceThe article presents a novel approach to automatic recommendation and summarization of bite-sized or 'snackable' media content to accommodate the increasing demand and volume of audio and video clips. It emphasizes the integration of video action recognition and viewership analysis to curate personalized teasers and highlights efficiently. The work contributes to the advancement of content consumption by aiding the selection process in an era where short-form content is becoming increasingly prevalent.
2021-02-18 - Scaling New Peaks: A Viewership-centric Approach to Automated Content Curation; Subhabrata Majumdar, Deirdre Paul, Eric Zavesky; Conference on Statistical Practice, 2021 [PDF]video summarization, viewership data, anomaly detection, personalized content, big data, automated curation, content highlights, video streaming, viewership metrics, content metadataThe article introduces an automated method for video content summarization that leverages viewership data to identify segments of high interest without manual curation. The method employs statistical anomaly detection on viewership metrics to find 'seed' segments, which are enhanced with content metadata and personalized for the final summary. Two case studies demonstrate the method's flexibility and effectiveness when compared to traditional highlight videos, by analyzing the US Democratic Presidential Debate and Wimbledon Women's Final from 2019.
2020-11-13 - Smartmedia: Locally & Contextually-Adapted Streaming Media; Yaron Kanza, David Gibbon, Divesh Srivastava, Valerie Yip, Eric Zavesky; SIGSPATIAL '20: 28th International Conference on Advances in Geographic Information Systems [PDF]streaming media, adaptive streaming, real-time content modification, geoblocking, geofencing, context-aware services, spatiotemporal information, media personalization, copyright protection, password sharing preventionThe article presents a novel streaming media technology that dynamically adjusts the content delivered to users based on spatiotemporal context, including location, time, and nearby individuals, enabling contextually relevant and appropriate media experiences. It introduces the vision of adaptive, responsive, and transposable streaming media, where, akin to web apps reacting to user interactions, streamed content can be modified in real-time to suit the viewer's environment. The technology also addresses issues like copyright protection through hyperlocal geoblocking and offers non-intrusive solutions to prevent password sharing by adapting media access control based on context.
2019-11-07 - Machine Learning and AI: Applications, Legal and Ethical Issues and Practical Guidance for Prosecution; Umesh M. Desai, Ed Guntin, Bruce E. Stuckman; University of Texas Advanced Patent Law, October 2019 (panel)Machine Learning, AI Applications, Ethical Challenges, Legal Issues, Prosecution Guidance, Marketplace Models, Personalization, Bias, USPTO, Alice Corp DecisionThe article discusses the tension between AI's market applications and the legal and ethical issues arising from AI-generated insights and models, emphasizing the need for clarity in ownership and responsibility. It explores the marketplace for self-generated AI models, called "Driverless AI" or "AutoML," highlighting challenges in ensuring quality, safety, and ethical appropriateness. Furthermore, it offers practical guidance for navigating the patenting process in the USPTO for AI-related inventions, advising on how to frame AI concepts to avoid rejections under the Alice Corp decision, such as linking AI to specific technical solutions and avoiding language that implies human mental processes.
2019-07-08 - END TO END STORY SEGMENTATION WITH HIERARCHICAL GRU-CNN MODEL ; Zhipeng Fan, Eric Zavesky, Zhu Liu; unpublishedstory segmentation, bi-directional GRU, attention mechanism, 2D convolutional neural network, end-to-end model, text classification, machine learning, NLP, closed captions, TDT2 datasetThe proposed model introduces bi-directional GRU with an attention mechanism and employs a novel 2D convolutional neural network, rather than traditional 1D convolution, to extract and summarize features for story segmentation in TV News closed captions. The model is designed to perform in an end-to-end manner, which simplifies the story segmentation task as a classification problem and eschews the need for hybrid neural network and graphical model approaches. It demonstrably surpasses the previous state-of-the-art on the TDT2 dataset, indicating the effectiveness of the 2D CNN and the classifier-only approach with appropriate post-processing.
2018-11-06 - OUR-COS : Opt-in UseR Context and Ontological Service
An Opportunistic, Omni-Channel Service and Advertisement Platform
; Eric Zavesky; unpublished
OUR-COS, user engagement, privacy-sensitive inputs, contextual services, evolving ontology, omni-channel aggregation, opportunistic service delivery, user context, service personalization, advertisementThe paper presents OUR-COS, an innovative architecture that integrates privacy-sensitive inputs such as touch, speech, gestures, and gaze with contextual services to enhance user engagement. It advocates for an evolving contextual ontology that aligns with the varied expectations of humans, digital interfaces, and service providers, ensuring cohesive service and advertisement experiences. The proposed system leverages omni-channel service aggregation and opportunistic service delivery tailored to user contexts, improving relevance and functionality through a flexible, user-centric approach.
2018-07-10 - A Stitch in Time - Autonomous Model Management via Reinforcement Learning; Elad Liebman; Eric Zavesky; Peter Stone; Proceedings of the 17th International Conference on Autonomous Agents and Multiagent Systems (AAMAS), July 2018. [PDF]Concept Drift, Autonomous Systems, Reinforcement Learning, Markov Decision Process, Model Management, Model Retraining, Sequential Decision Making, Data Distribution, Meta-Learning, Drift Detection MethodsThe article introduces a novel reinforcement learning (RL) framework for autonomous model management in response to concept drift, which is modeled as a sequential decision-making problem using Markov Decision Processes. This framework autonomously learns a policy for adapting models to changes in data distribution, focusing on the trade-off between learning costs and overall performance, rather than on specific underlying models or data. The proposed method outperforms existing meta-learning policies and drift detection methods by leveraging distributional properties of data for decision making, as opposed to relying on error rates or fixed parameters.
2018-02-12 - CARS: Collaborative Augmented Reality for Socialization; Wenxiao Zhang, Bo Han, Pan Hui, Vijay Gopalakrishnan, Eric Zavesky, Feng Qian
; HotMobile '18: Proceedings of the 19th International Workshop on Mobile Computing Systems & Applications,February 2018,Pages 25–30 [PDF]
Augmented Reality, Collaborative Framework, Quality of Experience, End-to-End Latency, Cloud-Based AR, Device-to-Device Communication, Computation Resource Reuse, User-Perceived Latency, Immersive Experience, Networking Resource Efficiency.The CARS framework improves user-perceived Quality of Experience in AR by leveraging the social interactions of proximate users to collaboratively share computation results, reducing end-to-end latency. It enables the reuse of networking and computational resources through intelligent sharing of information, allowing for a more immersive AR experience across both cloud-based and local environments. CARS significantly decreases object-recognition latency by about 40% compared to standard cloud-based AR by facilitating Device-to-Device communications and local processing based on cloud-obtained results.
2017-05-01 - Do Not Capture: Automated Obscurity for Pervasive Imaging; Moo-Ryong Ra;Seungjoon Lee;Emiliano Miluzzo;Eric Zavesky; IEEE Internet Computing
IEEE, IEEE Internet Computing, vol. 21, no. 3, pp. 82-87, May-June 2017, .
Do Not Capture, privacy, pervasive imaging, automated obscurity, mobile devices, face detection, motion detection, distributed system, real-time blurring, computer vision algorithmsDo Not Capture (DNC) is an innovative technology that protects individual privacy in public spaces by allowing individuals to become obscured in real-time from photos and videos captured by others' mobile devices, using a combination of face and motion detection algorithms. To enable this, the subject's device communicates a feature vector to the device capturing the media; if the subject is detected in the frame, their face is blurred before the media is saved or shared. The DNC technology is designed to be fully distributed, requiring no cloud support and offers up to 93.75 percent accuracy, even when tested in dynamic indoor and outdoor environments with real Android smartphones.
2014-11-03 - Hierarchical Feature Learning using Sparse Coding for Facial Semantic Analysis; ; unpublishedfacial semantic analysis, sparse coding, feature learning, deep neural networks, facial landmarks, multi-class classification, generative model, hierarchical feature learning, support vector machines, automatic frameworkThe proposed framework introduces a novel multistage architecture for facial semantic analysis using hierarchical feature learning with sparse coding that is fully automatic and generative, eliminating the need for labeled data during feature extraction. By learning features at facial landmark locations, the system effectively captures local facial structures to handle variations in pose and expression, which is robust to resolution changes and allows the same feature set to be used across different semantic classification tasks. The integration of sparse coding and deep neural networks to simultaneously analyze multiple facial semantics, including expression, age class, and gender, is a distinguishing and innovative aspect of this work.
2014-03-01 - AT&T Research at TRECVID 2013: Surveillance Event Detection; Xiaodong Yang; Zhu Liu; Eric Zavesky; David Gibbon; Behzad Shahraray; YingLi Tian; 2013 TREC Video Retrieval Evaluation [PDF]surveillance event detection, Space-Time Interest Points, Dense Trajectories, Fisher Vector, CascadeSVMs, feature-level fusion, decision-level fusion, imbalanced data, interactive detection environments, TRECVIDThe surveillance event detection system combines spatio-temporal feature extraction methods, namely Space-Time Interest Points (STIP) and Dense Trajectories (DT), with Fisher Vector encoding to robustly characterize human actions in videos. A CascadeSVMs algorithm addresses the challenge of data imbalance in surveillance footage by adapting to specific events and camera views. The system enhances detection accuracy through feature and decision-level fusions and employs two interactive environments for evaluating event detection performance.
2013-10-02 - Appearance, context and co-occurrence ensembles for identity recognition in personal photo collections; Archana Sapkota;Raghuraman Gopalan;Eric Zavesky;T. E. Boult; 2013 IEEE Sixth International Conference on Biometrics: Theory, Applications and Systems (BTAS), Arlington, VA, USA, 2013 [PDF]face recognition, ensemble approach, visual context, social co-occurrence, feature fusion, consumer photo collections, LFW dataset, Gallagher collection, unmodeled correlations, classification accuracyThe technical novelty in the article lies in proposing an ensemble-based face recognition approach that systematically combines diverse information sources (appearance, visual context, and social co-occurrence) to address the issue of unmodeled correlations in face data for consumer photo collections. The approach shows a 2.9% performance increase in verification accuracy on the LFW dataset over previous methods without requiring additional training data. Additionally, the new method achieves over 17.5% improvement in face identification tasks on the Gallagher personal photo collection through the utilization of visual context and social feature ensembles.
2013-08-26 - On the improvement of multimodal voice activity detection.; Matthew Burlick; Dimitrios Dimitriadis; Eric Zavesky; Proc. Interspeech 2013, 685-689, [PDF]Voice Activity Detection, Multi-Modal Fusion, Audio-Visual Integration, Feature Fusion, Decision Fusion, Multi-Tier Classification, Robustness to Noise, Computer Vision, Machine Learning, Speech RecognitionThis article presents an advanced audio-visual voice activity detection (AV-VAD) system that employs both feature fusion and decision fusion with a multi-tier classification strategy, improving robustness against noisy conditions. The novel system integrates a large array of features from both acoustic and visual modalities, using the outputs of base classifiers, as well as raw features, for training a subsequent higher-order classifier. Experimental results show significant improvements in voice activity detection accuracy on CUAVE and MOBIO datasets over traditional approaches, demonstrating the efficacy of the system's dynamic fusion technique.
2013-07-01 - Joint Audio-Visual Processing for Video Copy Detection; Zhu Liu; Eric Zavesky; David Gibbon; Behzad Shahraray; Academic Press Library in Signal Processing
Volume 5, 2014, Pages 417-455
video copy detection, hash-based indexing, multimedia databases, content identification, TRECVID, computer vision, audio-visual information, large-scale search, mobile applications, content transformationThe article introduces a scalable video copy detection system that leverages both audio and visual information, utilizing hash-based indexing and search to efficiently link new content to known content amidst various transformations. The novel approach combines proven computer vision techniques with robust indexing strategies, enabling quick identification of content for consumer and commercial applications, as demonstrated by the system's strong performance in TRECVID tasks. This work showcases the practical implementation of multimedia database algorithms in real-world applications, including mobile retrieval and large-scale video content library management.
2013-02-03 - VidCat: An Image and Video Analysis Service for Personal Media Management; Lee Begeja; Eric Zavesky; Zhu Liu; David Gibbon; Raghuraman Gopalan; Behzad Shahraray; IS&T/SPIE International Conference on Multimedia Content Access: Algorithms and Systems VII [PDF]cloud storage, personal media management, content analysis, face recognition, visual content similarity, video segmentation, near-duplicate detection, API, mobile platforms, automated tagging.VidCat introduces a cloud-based personal media management system that enhances the organization, tagging, and retrieval of photos and videos using face recognition, visual content similarity, time, and date metadata. It incorporates advanced content analysis algorithms for video segmentation, near-duplicate detection, and face detection, and evaluates its effectiveness using standard datasets. The system offers a client-accessible interface via HTTP-based APIs, with demonstrated utility in a mobile environment for both iOS and Android platforms, enabling straightforward media uploading and management.
2012-02-07 - Large Scale Multimedia Search on Mobile Devices; Zhu Liu; Eric Zavesky; Behzad Shahraray; The First Multimedia and Vision Meeting in Greater New York Area (talk)multimedia search, mobile devices, content recognition, minimal user input, real-time retrieval, network latency, resource footprint, accuracy, content sources, platform compatibilityThe technical novelty of the article lies in the development of a multimedia search system that enables mobile devices to use their built-in sensors to identify and retrieve relevant content from various databases without necessitating direct user input. The system is designed to provide quick and accurate content recognition that satisfies immediate user information needs while maintaining a low resource footprint and minimal network latency. Additionally, it ensures compatibility across diverse content sources without the need for specialized platform software.
2012-01-15 - Combining Content Analysis of Television Programs with Audience Measurement; David Gibbon; Zhu Liu; Eric Zavesky; Deirdre Paul; Deborah Swayne; Rittwik Jana; Behzad Shahraray; IEEE Consumer Communications and Networking Conference [PDF]IPTV, content analysis, audience metrics, media segmentation, automated systems, commercial detection, topic segmentation, electronic program guide, metadata, infrastructure planningThis work introduces a system that combines IPTV audience metrics with content-based analysis for a more granular understanding of viewer reactions to programs and advertisements, verified against news programs. It provides an automated methodology to detect commercial breaks and topic transitions, enhancing electronic program guide metadata with detailed, time-indexed content segmentation. The system enables broadcasters and advertisers to correlate audience tuneaway events with specific media content segments, offering insights for content delivery network management and infrastructure planning.
2011-12-07 - AT&T Research at TRECVID 2011; Zhu Liu; Eric Zavesky; Ning Zhou; Behzad Shahraray; 2011 TREC Video Retrieval Evaluation [PDF]content-based copy detection, instance-based search, TRECVID 2011, picture-in-picture detection, audio features, object-level detection, video retrieval, integer-score reranking, automated search, evaluation resultsThe technical novelty of the article involves enhancements to AT&T's content-based copy detection system with a focus on speed and the addition of a picture-in-picture detector alongside alternative audio features. It explores the application of the CCD system to the novel instance-based search task, which involves detailed object-level copy detection and introduces integer-score result reranking for video segments. The paper reports how these advancements performed in the TRECVID 2011 evaluations, specifically for content-based copy detection and instance-based search tasks in video retrieval.
2011-07-11 - LipActs: Efficient Representations For Visual Speakers; Eric Zavesky; 2011 IEEE International Conference on Multimedia and Expo [PDF]LipActs, histogram of oriented gradients (HOG), speaker retrieval, speaker verification, feature extraction, temporal sampling, spatial partitioning, visual speech recognition, biometrics, equal error rate (EER).The article presents LipActs, a novel method for visual speaker retrieval and verification without relying on complex tracking systems, by using modern histogram of oriented gradients (HOG) alongside spatial and temporal quantization techniques. The approach features improved accuracy through multi-resolution spatial quantization and pooled temporal frames, significantly enhancing performance over previous methods. The study demonstrates the effectiveness of LipActs across datasets from YouTube and forward-facing mobile cameras, highlighting a 37-73% improvement in the equal error rate (EER) for visual speaker recognition.
2010-11-15 - AT&T RESEARCH AT TRECVID 2010; Zhu Liu; Eric Zavesky; Behzad Shahraray; Neela Sawant; 2010 TREC Video Retrieval Evaluation [PDF]Content-Based Copy Detection, Instance-Based Search, TRECVID, Audio Features, Video Retrieval, Automated Search, Visual Object Queries, Digital Video Indexing, Heterogeneous Dataset, Multimedia AnalysisThe technical novelty of the published article lies in improving the Content-Based Copy Detection (CCD) system from TRECVID 2009 by incorporating audio features and enhancing its efficiency and scalability. The paper also introduces an automated Instance-Based Search (INS) system for TRECVID 2010, which leverages audio, video, and textual cues to search for specific instances within a video dataset. This work addresses the challenge of instance-based search through the heterogeneous IACC.1 dataset, with a particular emphasis on identifying similar instances based on visual object queries alongside minimal textual descriptions.
2010-10-29 - Unsupervised Event Segmentation of News Content with Multimodal Cues; Eric Zavesky; Andrea Basso; Mattia Broilo; Francesco G.B. De Natale; ACM Multimedia AIEMPro’10, October 29, 2010, Firenze, Italy [PDF]Unsupervised Clustering, Content Segmentation, Multimedia Content, Semantic Concepts, Video Retrieval, Feature Extraction, Coherence Rules, Media Consumption, Event Detection, Content Structure ExtractionThe article presents a novel unsupervised method to automatically extract structured content from news videos by clustering independent features such as audio, visual appearance, and mid-level semantic concepts using coherence rules to create coherent multimedia events. Unlike traditional methods, this system does not rely on pre-learned models or comprehensive training, allowing it to adapt to diverse and varying content without the constraints of previous supervised techniques. It also evaluates the impact of segmentation inaccuracies on video retrieval and content consumption, acknowledging the changing paradigms of media consumption to shorter, dynamic segments.
2008-10-30 - CuZero: Embracing the Frontier of Interactive Visual Search for Informed Users; Eric Zavesky, Shih-Fu Chang; MIR '08: Proceedings of the 1st ACM international conference on Multimedia information retrievalOctober 2008Pages 237–244 [PDF]CuZero, Visual Search, Query Formulation, Interactive Search, Real-Time Recommendations, User Interface, Information Retrieval, Concept Space Navigation, Visualization, Search EfficiencyCuZero addresses the shortcomings of visual search systems by incorporating real-time recommendations of visual concepts as the user types each word, leveraging lexical mapping, statistical occurrence, and search result mining. The system introduces a novel visualization interface for seamless navigation through the concept space, allowing users to explore multiple query permutations without time-consuming reformulation. CuZero creates an environment where users can rapidly review various query combinations and navigate different types of queries (semantic concepts, visual descriptors, example content) within a single session, enhancing the search experience and efficiency.
2008-10-26 - Columbia University/VIREO-cityU/IRIT TRECVID2008 high-level feature extraction and interactive video search; Chong-Wah Ngo, Shih-Fu Chang, Junfeng He, Yu-Gang Jiang, Elie El Khoury, Akira Yanagawa, Eric Zavesky; 2008 TREC Video Retrieval Evaluation [PDF]High-Level Feature detection, local features, global features, mean average precision, concept detection, cross-domain learning, multimedia search, interactive search task, automated recommendation, re-ranking strategiesThe technical novelty lies in demonstrating that local features alone achieve competitive results in High-Level Feature (HLF) detection tasks without the need for additional web images or specialized features like audio/face recognition, with a mean average precision (MAP) of 0.157. The integration of local and global features slightly enhances performance (MAP 0.162), while methods to incorporate web images and address domain differences show improvement for concept detection with limited training data. The interactive video search system, CuZero, successfully integrates user feedback with automated recommendation techniques to outperform standard manual search methods, with further performance gains observed through low-computation re-ranking strategies.
2008-10-12 - Cross-domain learning methods for high-level visual concept classification; Wei Jiang, Eric Zavesky, Shih-Fu Chang, Alex Loui; 2008 15th IEEE International Conference on Image Processing [PDF]Cross-Domain SVM, CDSVM, support vectors, domain adaptation, classification, machine learning, semantic concept classification, feature distribution, TRECVID benchmark, precision improvementThe article describes the development of a Cross-Domain SVM (CDSVM) method to transfer support vector knowledge from one domain to facilitate classification in another, enabling enhanced precision without significant computational overhead. It includes an extensive review and comparison of contemporary SVM-based cross-domain learning techniques and introduces a selection criterion to determine the most suitable cross-domain method for specific concepts. The CDSVM demonstrated a 7.5% increase in mean average precision across 36 concepts when evaluated using the TRECVID benchmark data set.
2008-07-08 - Columbia University's semantic video search engine 2008; Eric Zavesky, Shih-Fu Chang; CIVR '08: Proceedings of the 2008 international conference on Content-based image and video retrieval [PDF]cuvid, video search engine, semantic search, LSCOM, concept classifiers, real-time system, multimedia ontology, user interaction, automatic revision, TRECVIDThe paper advances 'cuvid', Columbia University's video search engine, by incorporating real-time interaction with large video datasets and automatic concept classifier adjustments drawn from the LSCOM ontology for enhanced semantic searches. The system notably reduces the need for users to understand the full lexicon, as it instantaneously revises the importance of concepts based on user interactions and search results. This evolution represents a significant step in large-scale video search engines, with user-driven concept influence adjustments enhancing the relevancy and precision of search outcomes.
2008-07-07 - Visual islands: intuitive browsing of visual search results; Eric Zavesky, Shih-Fu Chang, Cheng-Chih Yang; Proceedings of the 7th ACM International Conference on Image and Video Retrieval, CIVR 2008, Niagara Falls, Canada, July 7-9, 2008 [PDF]visual islands, multimedia exploration, feature projection, interactive browsing, image search, dimensionality reduction, user interface, query rank, result visualization, display configurationThe "visual islands" framework presents an innovative method for interactive exploration of digital multimedia by reorganizing search results or photo collections using a localized, non-global feature projection to generate 2D display coordinates, facilitating fast browsing and intuitive understanding. This approach maintains the relevance of the original query rank while allowing for non-linear exploration, and leverages a page-level analysis over a traditional global analysis, resulting in faster and more dynamic computing of display layouts. The prototype system demonstrates significant improvements in user comprehensibility and precision in navigating large image datasets compared to conventional methods.
2008-06-26 - Brief and High-Interest Summary Generation: Evaluating the AT&T Labs BBC Summarizations; Zhu Liu; Eric Zavesky; Behzad Shahraray; David Gibbon; Andrea Basso; TVS '08: Proceedings of the 2nd ACM TRECVid Video Summarization WorkshopOctober 2008 [PDF]Video summarization, TRECVID, Clustering, Shot segmentation, Salience, Redundancy elimination, Frame importance, Content variation, Playback configurations, Rushes videoThis article introduces a video summarization system designed for rushes summarization that segments videos into shots, applies multi-stage clustering to eliminate similarity, and employs a frame importance valuation based on temporal variation and spatial salience. The system is tested with dynamic and native playback configurations to balance ground truth inclusion and browsing ease, and is evaluated using the TRECVID 2008 benchmark where it achieves a high inclusion rate and comprehensibility. The technical novelty stems from the integration of content-based sampling and salience metrics in video summarization for efficiently processing videos with redundant content, such as rushes.
2007-10-08 - AT&T RESEARCH AT TRECVID 2007; Z. Liu, E. Zavesky, D. Gibbon, B. Shahraray, P. Haffner; 2007 TREC Video Retrieval Evaluation [PDF]shot boundary detection, rushes summarization, support vector machines, visual features, human-centric video summary, speed optimization, algorithm efficiency, pattern recognition, multimodal analysis, TRECVIDThe technical novelty of the AT&T system includes the introduction of new visual features for enhanced detection of cuts, dissolves, and fast dissolves in shot boundary detection (SBD), the implementation of a support vector machine (SVM) based verification method to improve accuracy and robustness in SBD, and an innovative multimodal approach for rushes summarization that incorporates both speech and face information to generate human-centric summaries with minimal redundancy.
2007-10-08 - Columbia University TRECVID 2007 High-Level Feature Extraction; S. Chang, W. Jiang, A. Yanagawa, E. Zavesky; 2007 TREC Video Retrieval Evaluation [PDF]Cross-Domain SVM, CDSVM, high-level feature classification, support vector machine, domain adaptation, TRECVID, machine learning, comparative study, selection criterion, performance improvementThe article introduces a Cross-Domain SVM (CDSVM) algorithm for improving the performance of high-level feature classification in videos by adapting support vectors from one domain for use in another. A comparative study of current SVM-based cross-domain learning methods is conducted, using TRECVID datasets for evaluation. The authors propose a novel selection criterion for determining the most effective cross-domain learning method to use under various conditions, leading to a reported 22% improvement in classification performance for certain concepts.
2007-09-18 - Searching Visual Semantic Spaces with Concept Filters; Eric Zavesky, Zhu Liu, David Gibbon, Behzad Shahraray; ICSC '07: Proceedings of the International Conference on Semantic Computing [PDF]semantic concepts, image search, video data exploration, LSCOM lexicon, user interaction, semantic classifiers, content-based search, image database, broadcast television, concept space navigationThe proposed system enhances image search by incorporating 374 semantic concepts from the LSCOM lexicon to enable rapid exploration of vast video data sets, improving the intuitiveness and efficacy of user interactions. By utilizing a combination of traditional image search with a semantic classifier-based approach, the system allows users to navigate quickly through a large and diverse image database sourced from broadcast television. The system's innovation focuses on leveraging the semantic classifiers' scores to facilitate user interaction with the automatically organized concept space, rather than developing new classifier methods.
2007-07-09 - Columbia University's semantic video search engine 2007; Shih-Fu Chang, Lyndon S. Kennedy, Eric Zavesky; Proceedings of the 6th ACM International Conference on Image and Video Retrieval, CIVR 2007, Amsterdam, The Netherlands [PDF]Video Search Engine, CuVid, Semantic Search, Multimodal Search, Visual Concept Detectors, Broadcast News, Keyword Search, Image Search, Near-Duplicate Search, User InterfaceColumbia University's video search engine 'CuVid' facilitates semantic and multimodal search within broadcast news video collections, employing 374 visual concept detectors for enhanced querying. The system compares the efficacy of different individual and combined modalities, including keyword, image, near-duplicate, and semantic concept searches, requiring only a simple web interface for user interaction. This search platform is significant in its ability to integrate various search methodologies and provide a user-friendly means for in-depth video analysis without the need for complex tools."
2006-10-09 - A Fast, Comprehensive Shot Boundary Determination System; Zhu Liu; David Gibbon; Eric Zavesky; Behzad Shahraray; Patrick Haffner; 2007 IEEE International Conference on Multimedia and Expo [PDF]shot boundary determination, finite state machine, support vector machine, TRECVID, visual feature extraction, SBD system, two-dimensional motion compensation, algorithm modularization, video retrieval, video processingThe technical novelty in the discussed SBD algorithm stems from the use of multiple finite state machine (FSM) based detectors specialized for different types of shot boundaries, alongside the integration of support vector machines (SVM) for enhanced cut and dissolve detection. Advances include the application of two-dimensional motion compensation, color information utilization, and a modular architecture enabling independent adjustment and easy expansion with additional detectors for new shot boundary types. The system demonstrated superior results in the TRECVID 2006 evaluation, ranking highest in overall performance for shot boundary determination.
2005-10-28 - Columbia University TRECVID-2005 Video Search and High-Level Feature Extraction; Shih-Fu Chang, Winston Hsu, Lyndon Kennedy, Lexing Xie, Akira Yanagawa, Eric Zavesky, Dong-Qing Zhang; 2005 TREC Video Retrieval Evaluation [PDF]High-Level Feature Extraction, TRECVID, parts-based representation, statistical models, computer vision, search task, audio/visual cues, concept search, near duplicate detection, story segmentationThe article reports on novel parts-based statistical approaches for High-Level Feature Extraction in TRECVID 2005, which were applied uniformly across 10 concepts to ascertain which high-level features are most enhanced by this methodology. Novel methods in the Search Task combined audio/visual cues, concept search, near duplicate detection, and story segmentation to improve video search capabilities, with significant improvement demonstrated across specific topics. Comparative analysis and a high-level performance measurement, using Mean Average Precision (MAP) and analysis tools, are employed to evaluate the effectiveness of these techniques against traditional ones.
an illustration of a computer and browser running copy writing software on the screen with swirling connections to data, photo events, and pie chart visualizations outside of the frame a colorful background composed of overlapping welcome messages spread out on a desk written in different mediums like scrolls, printed emails, carrier pigeon notes, flyers, fortune cookie wrappers


The publication of intellectual property codifies both works in progress as well as a future of things to be, for both defensive and creative claims. Below are selected works with articles documenting eventual fuition and a verbose list career filings.

Data parking within offline community system (US10798592B2)

Maintain data flow even when devices disconnect from the network. The system identifies offline devices within a community and uses a mobile aggregation device as a data lifeline, traveling to create a link with these isolated devices, collecting their data, and safely storing it. Upon return to a network-connected position, the system ensures that no data is lost, seamlessly preserving information in the face of unexpected connectivity issues.

Co-inventors: Jason Decuir, Robert Gratz, Eric Zavesky, James Pratt

data parking with US20200068419A1

UAV, drone, network intelligence, predictive scheduling, secure protocol buffering

Years after the AT&T filing, Apple has included an Emergency SOS mode. While it connects to satellites instead of UAVs, we assert the original filing provided a solution that was later brought to the consumer (link).

Network-controllable physical resources for vehicular transport system safety (US10885785B2)

Protect people with safety needs using technology to prevent accidents before they happen. This system employs processors to monitor network-connected vehicles and identify vulnerable individuals with registered safety concerns. Upon detecting a potential collision risk, it sends out urgent warnings to the vehicle in question and takes control of surrounding network-managed resources to mitigate the danger. This approach enhances safety and peace of mind in an increasingly connected world.

Co-inventors: Tan Xu, Eric Zavesky, Bernard S. Renger, Behzad Shahraray, David Crawford Gibbon, Zhu Liu

autonomous interactive safety US10885785B2

IoT, autonomous control, context-aware operation, interactive prremptive safety

In September 2022, Ford announced it is working on a mobile app designed to alert drivers to pedestrians and bicyclists nearby. (link)

Systems and methods for administrating suggested merchandising arrangements  (US11113350B2)

A method that personalizes shopping experiences like never before, analyzing product attributes with user behavior and language preferences. Utilizing advanced NLP/LLMs, it tailors product descriptions to match user history, making every recommendation strikingly relevant. Dynamic updates reflect user responses, continually enhancing the ability to present the perfect product description. This idea leans into personalized marketing and e-commerce interactions.

Co-inventors: Miranda Huet, James Pratt, Donald Perciful, Eric Zavesky, Michael Gonzales, Yupeng Jia

Dynamic text formulation

UAV, drone, network intelligence, predictive scheduling, secure protocol buffering

In September 2023, Amazon announced tools to automate the creation of product descriptions from only a set of image and text examples. (link)

Below is a searchable table of published patents with category tags and synopsis documents prepared by a GenAI analysis of the patent summary. Not all patents are in this form yet, so additional references are included below.

date / title / documentcategoriessynopsis
2023-05-30 / Content delivery system / US11665215B1 pdfpersonalized content delivery, quality of service metric, quality of experience metric, streamed content, content delivery network, user profile, adaptive streaming, remote learning, virtual classrooms, consumer engagementThe patent describes a system to personalize remote content delivery based on a user's experience and service quality, allowing for adaptive content streaming. It involves determining a quality of service metric and a quality of experience metric, then modifying streamed content for the consumer based on these metrics. This enhancement addresses the individual learning environment needs, accounting for factors like internet connection quality and user engagement.
2023-05-23 / System and method for contextual synthetic media generation / US11657852B1 pdfsynthetic media, deepfake, artificial intelligence, media manipulation, contextual analysis, media content modification, neural networks, content distribution control, deep learning, FaceSwapThe patent centers on a method for modifying media content by analyzing the content to determine the context of the event depicted, applying context-specific rules, and creating modified content to tailor distribution. This process includes the use of synthetic media generation techniques, like deep learning and artificial intelligence, to manipulate media while controlling its distribution to prevent consumer access. Embedding context awareness into media manipulation helps in preventing the spread of deceptive synthetic content, commonly known as deepfakes.
2023-01-03 / Cognitive persona embeddings for secure omni-channel personalized recommendations / US11546435B1 pdfmicro expressions, high-speed cameras, cognitive persona embeddings, personalized recommendations, behavioral analysis, user preferences, facial expressions, emotional response, omni-channel personalization, secure systemThe patent describes a system for capturing and analyzing micro expressions using high-speed cameras to provide personalized omni-channel recommendations or actions. The system integrates cognitive persona embeddings based on a user's emotional responses, which are difficult to conceal and occur involuntarily. The system utilizes an array of devices for interaction, thereby embedding user preferences and behaviors for secure personalization.
2023-05-30 / Dynamic avatars for customer support applications / US11663791B1 pdfdynamic avatars, customer support, emotional state, user preferences, workflow, personalized interaction, automated systems, extended reality, user satisfaction, customizationThis method involves creating dynamic avatars for customer support applications by identifying both a user's needs and emotional state, then customizing the avatar's features and presenting a stored workflow based on these parameters and the user's preferences. The avatar adapts not only to situational requirements but also to the emotions of the user, enhancing the interactive experience. The goal is to improve user satisfaction in automated support systems by offering a personalized and empathetic support interaction.
2023-01-31 / Detection and illumination of dark zones via collaborative lighting / US11570869B1 pdfdark zone detection, user profile, lighting customization, illumination control, internet of things, IoT, network-enabled light sources, adaptive lighting, user-centric lighting, wireless communicationThe patent describes a system that dynamically controls lighting based on detection of dark zones near a user, customizes the lighting according to user preferences stored in a profile, and sends instructions to network-enabled light sources to adjust the illumination to fit determined lighting features.
2023-02-28 / Enhanced emotive engagement with volumetric content / US11595636B1 pdfvolumetric video, emotion-based manipulation, contextual data, annotations, viewer feedback, actionable positions, immersive experience, machine learning, personalization, extended realityThe patent details a system for enhancing volumetric videos with annotations, contextual data, and viewer-specific actionable positions to manipulate the content for desired emotional impact. The system captures volumetric video, allows for viewer interaction through feedback and position data, and adjusts the presentation dynamically using machine learning. It aims to maximize user immersion by personalizing the volumetric experience based on the viewer's emotional response and viewpoint within the three-dimensional content.
2023-06-06 / Compositing non-immersive media content to generate an adaptable immersive content metaverse / US11671575B2 pdfimmersive content, metaverse, compositing media, user interaction, XR systems, virtual reality, augmented reality, media adaptation, real-world integration, virtual environmentThe patent describes a method for creating immersive content within a metaverse by compositing a user-provided media content with another, both depicting a subject, and adapting its presentation on a user's device based on the user's choices. This allows for the integration of real-world subjects into persistent virtual environments with the ability to modify presentation in response to user interaction. The technology addresses the application of XR systems by enhancing user experience with adaptable composite media in virtual spaces.
2022-08-02 / System and method for an adaptable, smart virtual assistant that supports long term care / US11402976B1 pdfvirtual assistant, normalized profile, anomaly detection, user interface, adaptation, user experience, personalization, dialogue system, smart virtual assistant, conversational agentThe invention details a virtual assistant with the capability to create a normalized user profile, detect anomalies that significantly deviate from this profile, and adapt its interface to accommodate these identified anomalies to improve user experience.
2023-01-31 / Augmented reality object manipulation / US11567572B1 pdfprocessing system, visual communication, object detection, action detection, notification, video analysis, user interaction, augmented reality, dynamic monitoring, feedbackA processing system identifies and compares actions related to objects within videos from two users in a visual communication session and notifies users of any differences in these actions. The system is designed for use with devices such as wearable augmented reality technology. This enables dynamic interaction monitoring and feedback during visual communication.
2022-12-20 / Method and system for augmenting presentation of media content using devices in an environment / US11533192B2 pdfmedia control package, Internet-of-Things, IoT, smart home, digital media content, playback synchronization, user experience, device control, timestamp, augmentationThe invention enables the creation of a media control package that allows Internet-of-Things (IoT) devices to augment the presentation of digital media content at specific timestamps. This media control package includes instructions that are utilized when the playback of the primary content reaches a particular timestamp, triggering the IoT devices to perform predefined actions. The aim is to enhance user experience by coordinating the state or behavior of smart home devices with on-screen events or contexts.
2022-10-04 / Virtual assistants for and conversations with non-human entities / US11463546B1 pdfvirtual assistant, non-human entities, communication, human user interaction, toys, animals, connected devices, networked environment, cross-entity, communications managementThe patent introduces a system where a virtual assistant uniquely represents non-human entities, enabling direct communication with a human user or indirect communication via another virtual assistant linked to the user. This system covers a range of non-human entities such as animals, toys, and connected devices, facilitating novel interactions and integrations in a networked environment. The patented technology describes methods and apparatuses for initiating and managing these cross-entity communications.
2022-05-31 / System for fan-based creation and composition of cross-franchise content / US11347387B1 pdfuser-generated content, franchise compliance, content creation, building blocks, composition module, search interface, processable instructions, policy enforcement, franchise objectives, content integrityThe patent discloses a system for managing fan-generated content within a franchise by allowing users to search, select, and manipulate building blocks extracted from the original content using a composition module while ensuring compliance with franchise policies. The system includes a processing unit and memory for storing executable instructions to assist users in creating content that aligns with the franchise's objectives without introducing anachronisms or misuse. The introduced mechanisms enforce content guidelines and maintain the integrity and coherence of the franchise while fostering user engagement and creativity.
2022-08-30 / System for addition and management of ad-hoc network attached compute (NAC) resources / US11431598B1 pdfnetwork attached compute, user equipment, location information, data analysis, trigger detection, thresholds, dynamic orchestration, network optimization, resource management, mobile applicationsThe invention involves a computer-implemented method for managing temporary network attached compute (NAC) apparatuses based on user equipment (UE) data, including their locations. It analyzes the UE data and NAC information to detect triggers, such as reaching certain thresholds. Upon detection, it orchestrates the dynamic movement and connection of the NACs to optimize network resources and services.
2022-03-22 / Providing adaptive asynchronous interactions in extended reality environments / US11282278B1 pdfextended reality, adaptive asynchronous interactions, avatar, persona selection, context-aware, user engagement, immersive technologies, dynamic representation, personalization, virtual environmentThe invention facilitates adaptive asynchronous interactions in extended reality (XR) environments by allowing a user's avatar to dynamically change personas based on surrounding conditions and interactions detected by the system. When another user attempts to interact with an avatar, the system evaluates the context and selects an appropriate avatar persona from a set of candidates. This dynamically tailored representation enhances user engagement in XR by providing context-aware personalization of avatar appearances.
2022-11-15 / Remote audience feedback mechanism / US11503090B2 pdfremote broadcast event, audience reaction estimation, audience grouping, representative selection, live feedback simulation, content delivery, event interaction, social distancing, user endpoint devices, media distributionThe patent introduces a method for delivering remote broadcast content from one user to multiple others, estimating reactions of the audience via data streams from their devices, grouping them according to their reactions, and then presenting the broadcast host with images of representative audience members for each reaction group. This technique simulates live audience feedback by using audience reaction analysis to enhance the remote event experience. It addresses the need for improved interaction between remote audiences and presenters, mimicking the feedback dynamics of in-person events.
2023-01-24 / Methods, systems, and devices for masking content to obfuscate an identity of a user of a mobile device / US11564085B2 pdfidentity protection, first responder, mobile network, user type, alternative content, identity concealment, mobile device, communications security, safety, network maskingThe invention involves a method for protecting the identity of first responders by determining their user type, identifying multiple identities based on that user type, and then providing alternative content to their mobile device to conceal their presence on a first responder mobile network from potential adversaries.
2022-10-25 / System and method for social immersive content rendering / US11483533B2 pdfimmersive media content, foreground rendering, background generation, point-of-view, synchronization, digital avatars, user interaction, spherical video, volumetric video, spatialized contentThe disclosed processing system improves the viewer experience in immersive media by creating and rendering foreground content, generating background content specific to a viewer's point-of-view (PoV), and synchronizing the display of both content types on the viewer’s device. The system also allows for social interaction in a shared immersive environment by synchronizing digital avatars with user activities such as gestures, voice, and gaze. Additionally, it facilitates presenting spatialized content adaptively from any viewpoint in both spherical and volumetric video formats, enhancing the level of engagement with the content.
2022-10-11 / External audio enhancement via situational detection models for wearable audio devices / US11470415B2 pdfwearable audio device, situational detection, audio adjustment, external audio data, sensor, microphone, volume control, processing system, earphone, modified audio dataThe patent details a processing system for a wearable audio device that detects specific situations using a situational detection model applied to audio captured by a microphone. Upon detecting a certain situation, the system adjusts the external audio data, such as altering the volume of the detected audio. These alterations are presented through the wearable's earphone to enhance the user's audio experience in varying contexts.
2022-11-08 / System for detection and analysis of viewership anomalies / US11496804B2 pdfreal-time viewership analysis, anomalous event detection, pattern recognition algorithm, second-by-second data logging, user interface, streaming, content ratings, audience behavior, programmatic advertising, standard viewership comparisonThe patent describes a system and method for real-time analysis of viewership data logged on a per-second basis to identify and analyze anomalous events during program streaming using pattern recognition algorithms, and then displaying the cause of such events through a user interface that contrasts these anomalies against a standard viewership pattern.
2023-06-13 / User-driven adaptation of immersive experiences / US11675419B2 pdfimmersive experience, media content customization, adaptive content, content variability, user engagement, real-time personalization, media distribution, narrative extraction, creator input, user experienceThis patent describes a method for creating customizable immersive experiences by analyzing and identifying variable details in media content components, allowing a creator to define variants to alter these details, and storing instructions for presenting these variants to the user, potentially replacing default segments for a personalized experience.
2021-10-19 / Adaptive immersive media rendering pipeline / US11151774B1 pdfimmersive environment, extended reality, XR systems, efficient rendering, user interaction, resource optimization, virtual object, computational load, multi-user support, dynamic optimizationThe patent introduces a method for efficient rendering in extended reality (XR) systems by detecting user interactions with virtual objects and rendering those objects in full detail, while other non-interacted objects are rendered in less detail, optimizing the use of rendering resources. This approach provides an immersive environment that is dynamically optimized based on user engagement, thereby reducing the computational load and enhancing the user experience. The system supports multiple users across various user devices, ensuring a consistent immersive experience within a managed resource ecosystem.
2022-10-25 / Time shifting extended reality media / US11481983B2 pdfextended reality, XR, time control, personal XR environment, time shift, user devices, individual experience, shared XR stream, independent navigation, non-disruptiveThe patent describes a method for individual time control in a shared extended reality (XR) experience, allowing a user to independently navigate to a different point in time in the XR stream without affecting the experience of other users. A personal XR environment is presented to the user's device in response to a time shift request, enabling personalized time navigation. The technology ensures that while one user revisits a prior segment of the XR stream, the overall shared experience for others remains uninterrupted.
2022-08-09 / System for summary segment association and dynamic selection for viewing with a content item of interest / US11412314B2 pdfcontent summarization, user context analysis, customized viewing, content delivery, video content, personalized recap, adaptive presentation, communication network, viewing time, context-awareThe patent describes a method for customizing and providing summary segments of video content to a user by analyzing the user's context and determining the most appropriate time and manner to present the content summary, based on the current viewing time. This method improves users' orientation to the content item of interest by adjusting the recapitulation segment to their specific needs. Additionally, the method engages with a communication network and devices to facilitate the adapted content delivery.
2022-08-02 / Variable-intensity immersion for extended reality media / US11405484B2 pdfextended reality, immersive experience, user engagement, adaptive content, user reaction monitoring, content variants, threshold deviation, personalized media, dynamic adaptation, baseline adjustmentThe patent presents a method for enhancing user engagement with extended reality media by dynamically adapting the immersive experience based on the user's reaction. It involves monitoring user reactions against a baseline and selecting from multiple variants of the media content to guide the user's reaction back towards the baseline. This adjustment is done when deviation from the baseline exceeds a predefined threshold, aiming to maintain an optimal engagement level.
2022-09-06 / System for nostalgic content links and playback / US11436640B2 pdfemotive response, content viewing, external media, nostalgic content discovery, viewer engagement, memory response, personalized media, content playback, emotional triggering, user experienceThe patent describes a system and method for enhancing a viewer's content experience by capturing their emotive response when viewing content for the first time and then detecting the emotive response upon subsequent viewings to trigger the retrieval and display of external media related to the captured emotions.
2021-12-07 / System for infinite windows with optical disparity and depth resolution / US11196944B2 pdfreal-time stitching, visual presentation, user viewpoint tracking, gaze direction, camera obstruction removal, realistic visual experience, indoor display, multiple cameras, augmented reality, live image processingThe patent discloses a system for generating an unobstructed visual presentation based on a user’s viewpoint and gaze direction, utilizing multiple real-time camera images that are stitched together to compensate for any obstructions observed by the cameras. By leveraging this technology, the system provides a continuous, realistic view of the outside from within a room without windows, enhancing the indoor experience. Real-time image processing allows for the inclusion of live elements and the removal of potential visual obstructions, offering an active and accurate representation of the external environment.
2022-04-19 / Immersion control system for extended reality systems / US11307651B2 pdfextended reality, XR, immersive experience, user experience level, immersion detail level, automatic adjustment, customization, XR equipment, tailored experience, communicationThe invention relates to tailoring extended reality (XR) immersive experiences by automatically adjusting the level of immersion based on a user's experience level, thereby customizing the experience to suit individual user capabilities or preferences. This adjustment is managed by determining a user experience level and then modifying the immersion detail level accordingly. The data that defines the tailored XR immersive experience is then communicated to the user's XR equipment.
2023-01-17 / Automated, user-driven, and personalized curation of short-form media segments / US11558650B2 pdfpersonalized media stream, user viewing habits, content curation, media segments, automated compilation, streaming services, media consumption behavior, user-driven personalization, dynamic assembly, media contentThe patent details a method for creating personalized media streams by selecting, modifying, and compiling media segments based on a user's viewing habits. It focuses on curating content that reflects individual preferences by adjusting media segments before they are integrated into a tailored stream. The process utilizes automation to ensure the final media product is both user-specific and dynamically assembled.
2023-01-24 / System for extended reality visual contributions / US11562818B2 pdfimmersive environment, extended reality (XR), user task, immersion objects, physical environment, virtual integration, tailored experience, XR device, task completion, user informationThis patent details a method for creating an immersive environment by integrating virtual immersive objects with a user's physical surroundings, tailored based on the task to be completed, user information, and the physical environment, and then communicating this tailored immersive environment to the user's XR device to enhance task completion.
2022-10-25 / Methods, systems, and devices for providing a user experience associated with a pause in presenting media content / US11483615B2 pdfmedia content, pause, communication device, playback, attributes, instructions, streaming, user experience, targeted content, media playback interruptionThis patent pertains to an advanced system for handling paused media content on a communication device by determining attributes associated with the pause and providing tailored instructions based on those attributes. The system can discern the reason for the pause and respond accordingly, enhancing the user experience during interruptions of media playback. This includes the possibility of displaying targeted content or information on the device during the pause state.
2022-07-19 / Apparatuses And Methods For Identifying Content Distribution Opportunities In Accordance With Advertising Performance / 11392983 pdfcontent distribution, advertising performance, dynamic distribution, real-time metrics, performance monitoring, threshold, additional outlets, automated selection, responsive content provision, content reach optimizationThe patent outlines a method for dynamically identifying and utilizing advertising performance data to determine when a piece of content should be distributed to additional outlets based on its performance exceeding a predetermined threshold. This enables responsive content distribution that leverages real-time advertising metrics, allowing content providers to quickly capitalize on successful content across various platforms. The technique involves monitoring content performance in one outlet, and upon reaching a success criterion, automatically selecting and providing the content to further outlets for enhanced reach and distribution.
2022-11-22 / Optimization of resource usage in cross-reality sessions / US11507424B2 pdfcross-reality, XR, resource optimization, user attention data, dynamic detail adjustment, computational efficiency, bandwidth management, real-time updates, immersive environments, performance enhancementThe patent introduces a system for optimizing resource usage in cross-reality (XR) sessions by dynamically adjusting the level of detail for entities based on user attention data, minimizing computing and bandwidth demands. Utilizing real-time attention tracking, the system provides resource-efficient updates to XR environments, enhancing performance and user experience. The technology addresses the challenges of computational and network resource constraints in rendering complex cross-reality environments by selectively updating elements within a user's focal point.
2022-08-09 / Cross-reality safety service / US11410518B2 pdfcross-reality safety, XR session monitoring, virtual environment, physical environment detection, safety issue detection, immersive technology, user alert system, environment modification, session management, risk mitigationThe patent centers around a cross-reality (XR) safety service that detects safety issues within XR sessions by monitoring the interplay between virtual and physical environments. If a safety concern is identified, the service responds by modifying the XR environment, ending the session, or alerting the user. This system aims to mitigate risks associated with the immersive nature of XR technologies, which may cause users to neglect the real-world surroundings.
2022-11-08 / Policy definition and enforcement for extended reality media sessions / US11494951B2 pdfextended reality, XR, user behavior policy, content monitoring, policy compliance, virtual reality, augmented reality, mixed reality, content modification, user experienceThe invention pertains to a method for rendering and managing an extended reality (XR) environment that enforces user behavior policies. It involves a processing system receiving a request to render the XR environment with a specified policy, monitoring user-contributed content for policy compliance, and modifying non-compliant content to achieve policy adherence. The system ensures the XR environment aligns with predefined user behavior policies, enhancing user experience and safety.
2023-01-17 / Techniques for real-time object creation in extended reality environments / US11556961B2 pdfextended reality, XR, user preferences, historical profile, local environment, sensor array, XR objects rendering, compute resources allocation, local compute resources, edge computingThe patent describes a method for selecting and rendering XR objects in an extended reality application personalized for a user, combining user preferences, historical profile data, and local environment information gathered by an array of sensors. It includes a resource allocation strategy that dynamically chooses between local and edge computing resources to facilitate the rendering process effectively. This personalized and resource-optimized rendering enhances the user experience in XR by adapting to both the user's context and technical system capabilities.
2021-10-05 / System for multi-presence interaction with extended reality objects / US11138803B1 pdfaugmented reality, collaborative interaction, local environment scanning, remote environment, user perspective, object characteristics, display device, interaction detection, control transfer, visual feed updateThe technical novelty involves an augmented reality system that allows a local user and a remote user to collaboratively interact with an object from their respective environments. The system provides customized image information to each user's display device, reflecting their unique viewpoints and interactions with the object. It enables seamless transfer of control over the object between the local and remote user, updating the visual feed accordingly.
2022-05-10 / Apparatus and method for providing content with multiplane image transcoding / US11330252B2 pdf3D video streaming, volumetric video, multiplane image (MPI), viewpoint information, adaptive-bitrate streaming, 6DoF, quality of experience, network resource optimization, computational efficiency, memory efficiencyThis patent discloses a system for efficient transmission and playback of 3D/volumetric video by utilizing a multiplane image (MPI) representation that responds to transmitted viewpoint information including 3D coordinates and timestamps, enabling adaptive streaming with reduced bandwidth and computational overhead. The disclosed technology addresses the challenges in streaming volumetric videos with six degrees of freedom (6DoF), by providing a solution that is computationally and memory efficient without compromising resolution for arbitrary topologies. The system also aims to integrate adaptive-bitrate streaming parameters to enhance the quality of experience for users.
2023-05-09 / Apparatuses and methods for managing tasks in accordance with alertness levels and thresholds / US11642059B2 pdfalertness monitoring, sensor data analysis, predictive alertness level, user notification system, cognitive burden mitigation, machine operation, fatigue management, safety, vehicular technology, proactive user assistanceThe invention focuses on monitoring a user's level of alertness using sensor data, predicting a required level of alertness for future operation of machinery, and delivering a timely notification before this future point to ensure the user can achieve the necessary alertness level. The system analyzes current alertness, compares it with the predicted need, and determines the optimal time to intervene with a notification to assist the user. This proactive approach addresses cognitive burdens and fatigue in decision-making processes by providing user assistance at a calculated time to maintain safety and efficiency in vehicular operations.
2022-08-16 / Apparatuses and methods for managing tasks in accordance with alertness levels and thresholds / US11412969B2 pdfalertness management, sensor data, cognitive load prediction, user engagement, vehicle operation, automated notification, safety, efficiency, fatigue mitigation, decision-making assistanceThis technology pertains to a system for managing user alertness by obtaining sensor data to assess a user's current alertness level, predicting the required alertness level for operating machinery at a future time, and providing a notification to the user to prepare them for the expected cognitive demand. The system is designed to predict cognitive load and optimize user engagement to maintain safe and efficient machine operation, such as in a vehicle. It includes the capability to determine when to alert the user in advance of a task to mitigate cognitive fatigue and ensure timely user responsiveness.
2021-06-22 / Apparatuses and methods for managing tasks in accordance with alertness levels and thresholds / US11039771B1 pdfuser alertness monitoring, sensor data analysis, predictive alertness levels, machinery operation readiness, cognitive burden mitigation, fatigue management, timely notifications, user engagement prediction, safety, efficient navigationThe invention pertains to monitoring user alertness levels utilizing sensor data, predicting required future alertness for operating machinery, and delivering timely notifications to ensure the user's readiness. It involves assessing current alertness, forecasting future alertness needed, and comparing these to determine when and what type of notification should be provided before the user needs to engage with the machine. The system is designed to mitigate cognitive burdens and response delays due to fatigue by providing preemptive notifications based on the anticipated demand for alertness.
2023-05-16 / Method for accelerating three-dimensional object segmentation with point cloud simplifications / US11651498B2 pdfpoint cloud, downsampling, machine-learning model, segmentation, object identification, data size reduction, computational efficiency, loss functions, Intersection-over-union, deep neural networksThe patent details a system that employs downsampling of a point cloud to reduce its data size, followed by the application of a machine-learning model, trained on the full point cloud, to perform segmentation and object identification on the downsampled point cloud, with the aim of reducing computational resource usage while maintaining accuracy.
2022-05-24 / Method for accelerating three-dimensional object segmentation with point cloud simplifications / US11341650B2 pdfdownsampling, point cloud, machine-learning model, segmentation, object identification, data size reduction, computational efficiency, memory overhead, training, accuracyThe invention involves a device featuring a processor and memory able to downsample a full point cloud to reduce its data size and use a trained machine-learning model to perform segmentation and object identification on the downsampled point cloud. This technique maintains segmentation accuracy while reducing computation and memory overhead associated with processing full point clouds. The machine-learning model is notably trained on the full point cloud to effectively label the downsampled point cloud.
2022-10-11 / Consistent generation of media elements across media / US11470404B2 pdfdigital model, media element, audiovisual media, rendering, metadata, insertion, scene, consistency, characteristics, databaseThe patent pertains to a method for ensuring consistent representation of media elements by retrieving their digital models from a database, rendering them with specific metadata that defines characteristics and limits, and then inserting the rendered element into an audiovisual scene. This process is designed to maintain the legal and artistic fidelity of the media element across different media productions. Metadata plays a crucial role in defining the renderable characteristics and boundaries to ensure repeated usage aligns with predefined constraints.
2023-05-23 / Methods, systems, and devices for identifying a portion of video content from a video game for a player or spectator / US11654372B2 pdfvideo game content, viewer feedback, popularity threshold, content distribution, esports, video game server, targeted advertising, content recording, viewer engagement, dynamic selectionThis patent describes a system for selecting and distributing portions of video game content based on viewer feedback. It assesses the popularity of video content displayed on viewers' devices and, when a popularity threshold is reached, selects additional related content for recording and distribution. The system's innovation lies in dynamically tailoring game content distribution to viewer preferences and engagement levels.
2022-01-04 / Methods, systems, and devices for identifying a portion of video content from a video game for a player or spectator / US11213758B2 pdfvideo content distribution, video game, viewer feedback, popularity threshold, content selection, subject matter identification, enhanced viewing experience, dynamic recording, esports, content presentationThe patent describes a system for selecting and distributing video content from a video game to viewers' devices, presenting the content, and collecting feedback for popularity assessment. If the content meets a predetermined popularity threshold, the system identifies and records a further segment associated with the same subject matter of the popular content. This dynamic content selection and distribution process is tailored based on viewer interaction and content popularity for enhanced viewing experiences.
2022-08-30 / System for ledger and parametric marketplace for algorithmic content creations / US11430420B2 pdfblockchain ledger, machine learning, media content, composition parameters, digital rights management, content analysis, user equipment, music streaming, content attribution, parameter recordingThe patent describes a device that uses machine learning to analyze media content for compositional parameters, records the parameters in a blockchain ledger distinguishing between original and existing works, and searches the ledger for similar works, reporting findings to the user.
2021-11-30 / Systems and methods for spatial remodeling in extended reality / US11189106B2 pdfmachine learning, remodeling proposals, decorating style, user feedback, 3D models, user engagement, design accuracy, rendered proposals, personalization, environment modelingThe patent details a machine learning system that generates remodeling proposals based on user's decorating style preferences and images or models of their environment, and refines proposals using user feedback. The system uses images or 2D/3D models of the desired remodeling space, processes them with machine learning to match a user’s style preference, and presents rendered proposals for review on user equipment. Subsequent iterations incorporate user feedback to refine and improve the design proposals, enhancing user engagement and design accuracy in the remodeling process.
2022-02-01 / Apparatus and method for deconflicting competing cross reality engagements / US11238835B2 pdfextended reality, XR, virtual object, conflict resolution, multi-user, communication session, context-aware, content presentation, shared XR space, multi-applicationThe invention provides a solution for managing and presenting virtual objects within an extended reality (XR) environment for multiple users, where it identifies conflicts when virtual or physical objects overlap and modifies the presentation of objects to resolve such conflicts. It tailors the display of virtual objects to each user's context during a multi-application communication session. Responsive adjustments prevent the obscuring of important content across different users' views within a shared XR space.
2021-07-20 / Evaluation of digital twins using social automatons / US11068046B1 pdfextended reality, digital twin, social automaton, simulation, virtual representation, immersive technology, interaction, environmental design, evaluation, XR mediaThis patent introduces a method for enhancing simulations by integrating social automatons into digital twins of physical environments, enabling the creation of extended reality media where virtual representations of individuals interact with the environment's digital counterpart, simulating real-life interactions. The social automatons are programmed with behaviors and characteristics authentic to the individuals they represent, enhancing the accuracy of interaction simulations within the XR framework. This allows for detailed evaluation of environmental designs or changes by visualizing their impacts on human interaction in various applications like industrial layouts, real-estate arrangements, and urban planning.
2022-07-19 / Content capture service / US11394875B2 pdfcontent capture application, capture suggestion, project definition, device capability, video content creation, media project assembly, camera functionality, user interface, network communication, content sharingThe patent describes a content capture application/service that provides capture suggestions to a user device based on a predefined project definition and the device's video and networking capabilities, resulting in media projects tailored to the intended viewing devices. It incorporates user device camera functionality, device-specific capabilities, and network attributes to optimize the creation and sharing of video content. The system guides users in capturing video content, assembles the media project, and facilitates output to designated recipients.
2023-01-17 / Cognitive stimulation in vehicles / US11557096B2 pdfinteractive content, automated vehicle, orchestrator component, augmented reality, virtual reality, vehicle action sequence, entertainment, user preferences, immersive experience, nearby vehicle notificationThe patent describes a system for enhancing the in-vehicle experience by providing interactive content, including augmented reality (AR) and virtual reality (VR), tailored to vehicle conditions, user preferences, and external information. An orchestrator component correlates vehicle travel routes with entertainment content and controls the vehicle to recreate action sequences from video programs, enhancing passenger immersion. It also includes mechanisms to notify nearby vehicles or invite them to participate when such sequences are being recreated.
2022-11-08 / Methods, systems, and devices for identifying viewed action of a live event and adjusting a group of resources to augment presentation of the action of the live event / US11496778B2 pdfvideo content aggregation, audience point of view, augmented reality, live event, content selection, streaming, video service provider, remote viewing, communication device, event capturingThe patent introduces a system for aggregating multiple video content streams of an event, selecting specific content based on a determined audience perspective, creating augmented reality content from this selection, and providing the AR content to users' communication devices for enhanced viewing experiences.
2021-11-02 / Methods, systems, and devices for identifying viewed action of a live event and adjusting a group of resources to augment presentation of the action of the live event / US11166050B2 pdfvideo content aggregation, augmented reality, point of view, audience members, event streaming, personalized presentation, communication device, content selection, live event, sensor informationThe technical novelty lies in the method of aggregating video content captured by audience members to determine a collective point of view, creating augmented reality (AR) content from the aggregated videos corresponding to the determined point of view, and then providing this AR content to individual communication devices for personalized presentation.
2022-08-30 / Method and apparatus for generating personalized payloads / US11430012B2 pdfcontent-based advertising, user context, purchase terms presentation, seamless user experience, advertising engagement, content flow maintenance, revenue optimization, ad suggestion algorithm, communication devices, transaction feedback loopThe innovation enables the selection of a first advertisement within content to trigger the presentation of purchase terms, followed by content and advertisement suggestions based on the initial content selection and the user's context. Upon completion of a purchase, a relevant second advertisement is selected and displayed to the user, maintaining engagement without interruption of the primary content flow. This creates a seamless advertising experience that can potentially increase revenue for content publishers and advertisers while minimizing user distraction.
2023-02-28 / Social communities assistant / US11593445B2 pdfsocial media, communication management, proxy system, single interface, networking sites, user updates, automated posting, social communities assistant, content sharing, multi-platformA single interface system serves as an intermediary to manage and post communications on multiple social networking sites on behalf of the user, addressing the issue of manually updating each site individually. The system supports various communication types including queries, responses, and notifications, operating as a unified proxy for the user across diverse social communities. This approach aims to alleviate the burden of managing multiple social network updates and prevents information from becoming stale due to infrequent user updates.
2022-09-06 / Property archivist enabled customer service / US11436611B2 pdfproperty archivist, automated customer service, electronic inventory, product detection, service request, problem diagnosis, automated purchasing, shipping, Internet of Things, smart devicesThe patent describes an apparatus and method for automated customer service via an electronic property archive which creates a product profile upon detection, initiates service requests upon problems, diagnoses issues, and facilitates automated purchasing and shipping of replacement parts.
2022-08-30 / Managing interactions of products and mobile devices / US11430016B2 pdfInternet of Things, IoT, sensors, user interest detection, mobile device, smart products, server communication, data retrieval, product information, location-basedThe disclosed technology relates to an IoT-based system that detects a user's level of interest in a product using sensors, and sends this data to a server to retrieve and display location-based information about a category of products on a mobile device. The system enhances user experience by facilitating the interaction between smart products and mobile devices to provide relevant information and control based on detected interest. It addresses existing technological limitations by creating a more interconnected and responsive environment for users within the IoT paradigm.
2022-07-05 / Systems and methods for applied machine cognition / US11380094B2 pdfserver, personality traits, inanimate objects, data evaluation, usage factors, interaction factors, environmental factors, personality assignment, avatar generation, internet of thingsThe patent describes a technological method whereby a server evaluates data from an object's interactions and environmental exposure over time to assign it human-like personality traits and an avatar. The personality is determined by comparing object data against specific personality-linked thresholds, which involve usage, interaction, and environmental factors. The resultant personality assignment is then utilized to generate a corresponding avatar that represents the inanimate object.
2023-04-04 / Detection of usage of a physical environment / US11620664B2 pdfauto-determination, physical environment usage, localization technology, intelligent actions, maintenance decision-making, usage patterns, virtual perimeters, asset management, location-based services, monitoring systemsThe patent discloses systems and methods for automatically determining partial usage of a physical environment utilizing localization technologies to detect the presence of a device or user. Intelligent actions are taken based on the detection, such as environment maintenance decision-making, influenced by single or cumulative usage patterns. The technology improves upon existing virtual perimeter applications like asset management, location-based services, and monitoring systems.
2022-08-16 / Detection of usage of a physical environment / US11416878B2 pdfauto-determination, physical environment usage, virtual perimeter, localization technologies, intelligent maintenance, environmental interaction, usage detection, geographical landscape, asset management, adaptive systemsThe patented technology relates to methods and systems for automatically determining partial usage of a physical environment by monitoring presence within a virtual perimeter and leveraging this data for environmental maintenance decisions. Unlike existing technologies, it enhances adaptability by intelligently recognizing not only presence but the extent of usage over time to guide specific actions. This innovation allows for optimized maintenance activities based on both singular and cumulative environmental interactions.
2022-04-12 / Assisted micro-environment interaction / US11304026B2 pdflocalization technology, user commonalities, virtual perimeter, environment interaction, wireless device location, dynamic adjustment, system, method, interaction management, user behaviorThe innovation presents advanced methods and systems for identifying commonalities among users within a specific area, using localization technologies. These commonalities can then be used to adapt the virtual perimeter for various applications, including enhanced interaction management within that environment. Improvements include the dynamic adjustment of virtual boundaries based on collective user attributes and behavior.
2022-09-06 / System and method for dynamic manipulation of content presentation / US11438395B2 pdfdynamic content manipulation, media content classification, user context determination, video complexity, audio characteristics, playback speed adjustment, automated media presentation, personalized consumption, mobile platforms, viewing experienceThe invention provides a system for dynamic manipulation of media content presentation by classifying content based on video complexity and audio characteristics, determining the user's context, and then applying manipulations like varying playback speed or repeating sections to the content based on this classification and user context. This allows for personalized media consumption that adapts to the viewer's available time and preferences, offering an enhanced viewing experience on mobile platforms. Unlike current solutions, this system automates content manipulation to optimally fit within the user's time constraints and consumption habits.
2022-10-11 / Determining power optimization for multiple radios based on historical power usage in advanced networks / US11470560B2 pdfreal-time power optimization, advanced networks, multiple radios, user equipment, power distribution, historical usage, performance results, location-based, applications, machine learningThe patent describes a method for real-time power optimization in advanced networks by determining and implementing a power distribution setting for devices with multiple radios, considering factors like historical radio power usage, performance results, current location, and ongoing applications. This power distribution is managed among radios of different types within user equipment. The system utilizes a combination of memory, processor, and possibly machine learning models to facilitate power optimization and reduce electromagnetic exposure.
2022-11-08 / Methods, systems, and devices for providing augmented reality content based on user engagement / US11496797B2 pdfaugmented reality, video content, points of interest, object identification, interactive viewing, content enhancement, communication network, user engagement, information retrieval, communication deviceThe patent describes a method for enhancing video content by identifying points of interest and objects within the video, obtaining relevant information about these objects, and generating augmented reality content based on this information to be presented on a user's communication device. This enables interactive viewing experiences where users can engage with augmented reality overlays that provide additional content related to the objects of interest within the video. The system utilizes a communication network to receive video content, process it for augmented reality enhancements, and deliver the augmented content to various devices.
2022-04-19 / Management of persistent network slices by a distributed learning system in a 5G or other next generation wireless network / US11310104B2 pdf5G, network slicing, resource allocation, scheduling, management, user device characteristics, historical data, artificial intelligence, machine learning, resource configurationThis invention pertains to an optimized system for the allocation, scheduling, and management of 5G network slice resources by employing a selection mechanism that configures resources based on user device characteristics and historical usage data. It also involves communicating the configured resources to network devices for subsequent allocation, ensuring proper resource distribution to the connected user devices as needed. Additionally, the system can include artificial intelligence/machine learning components to enhance the management of network slices.
2023-03-21 / Antenna selection and device orientation / US11611847B2 pdfantenna configuration, mobile device, automatic adjustment, performance optimization, historical data, signal strength, communication quality, device orientation, cooperative determination, wireless communicationThe invention enhances wireless device performance by implementing automatic antenna configuration and by adjusting the device's orientation or location based on historical performance data. The system facilitates proactive adjustments in mobile devices to optimize signal strength and communication quality. Cooperative determination of optimal configurations among multiple mobile devices is also featured to further improve overall performance.
2021-03-16 / Antenna selection and device orientation / US10952024B2 pdfwireless device, antenna selection, device configuration, orientation adjustment, location optimization, historical performance data, network performance, automatic configuration, mobile device, cooperative determinationThis patent describes a system for optimizing wireless device performance by automatically selecting antennas and adjusting mobile device configuration, which may include changes in orientation or location. The system leverages historical performance data to inform proactive adjustments for improved connectivity. Cooperative determination of optimal orientation or location between multiple devices is also covered to enhance overall network performance.
2021-03-30 / Apparatus and method for presentation of in-store visualizations and/or supply of products to customers / US10963918B1 pdfpersonalized shopping experience, shopper identification, custom visualization layout, multi-screen display, shoppers' route tracking, tailored visualizations, in-store navigation, dynamic content adaptation, movement detection, proximal display activationThe technical novelty lies in creating a personalized shopping experience by detecting a shopper's entry into a store, identifying the shopper, and generating a custom layout of visualizations to be displayed on multiple screens situated along the shopper's route within the store. As the shopper moves, the system tracks the movement and provides tailored visualizations on the proximal display screens, enhancing efficiency and personalization of the in-store experience. The individualized visualizations change as the shopper continues to different locations in the store, dynamically adapting the content to the shopper's journey.
2022-03-01 / Synchronizing multiple user devices in an immersive media environment using time-of-flight light patterns / US11265387B2 pdfsynchronizing devices, immersive media environment, time-of-flight, ToF, light patterns, fiducial detection, positional tracking, device orientation, augmented reality, virtual realityThe patent provides a method for synchronizing devices in an immersive media environment by detecting fiducials within time-of-flight light patterns projected by another device in the same environment. A device's location is determined by identifying the second device based on the fiducial's appearance and comparing the observed orientation of the fiducial with its expected orientation. This enables accurate positional tracking in immersive applications without requiring external sensors or cameras.
2021-06-01 / Traffic management system / US11024162B2 pdftraffic management system, sensor data analysis, policy agreement enforcement, visual indicators, autonomous vehicles, non-autonomous actors, real-time monitoring, policy violation notifications, community norms, visual indication identificationThe traffic management system identifies and addresses visual indication policy violations by processing sensor data, associating the identified visual indications with policy agreements, and providing notifications when a visual indicator system violates said policies. The system caters to environments with both autonomous and non-autonomous entities, such as various vehicles and pedestrians, ensuring adherence to traffic laws and community norms. It advances traffic management by integrating policy enforcement mechanisms with real-time sensor data analysis from the physical environment.
2022-10-11 / Application management of network slices with ledgers / US11470181B2 pdf5G network slicing, resource allocation, application-based scheduling, anomaly detection, AI/ML, blockchain, dynamic contracts, user equipment management, automated system, network resource managementThis patent centers on a method and system for automating the allocation and management of 5G network resources through dynamic contracts and scheduling based on users' applications requirements. Using an AI/ML component, the technology predicts and detects anomalous network slice usage, enabling rapid adjustments and potential expulsion of user equipment applications. It also integrates a blockchain component to facilitate secure and transparent application-based network slice requests and management.
2021-03-30 / Application management of network slices with ledgers / US10965777B2 pdf5G network, network slice management, contract-based resource allocation, user equipment, anomalous usage detection, multi-access edge computing (MEC), artificial intelligence (AI), machine learning (ML), resource scheduling, governing contractsThe patent describes a system for managing network slice resources in 5G networks that includes operations for receiving contract data, storing a governing contract to control resource access by applications on user equipment, and scheduling use of network slice resources based on the governing contract. The system aims to address the limitations of current network slice management in adapting to applications with varying requirements such as emergency services or high-bandwidth demand, and includes mechanisms for detecting and responding to anomalous usage. Advanced features such as multi-access edge computing (MEC) and artificial intelligence/machine learning (AI/ML) are integrated for efficient allocation, scheduling, and management of network slices.
2022-12-27 / Task delegation and cooperation for automated assistants / US11537417B2 pdftask delegation, automated assistants, sub-tasks, coordination, bots, artificial intelligence, task parsing, capability identification, proposal formulation, sequence mappingThe disclosed technology addresses the coordination of automated assistants (bots) by breaking down tasks into sub-tasks, identifying which bots are capable of performing each sub-task, and creating a sequence and mapping plan that dictates the order and allocation of sub-tasks to specific bots for efficient task completion.
2022-03-22 / Activation of extended reality actuators based on content analysis / US11282281B2 pdfextended reality, actuator synchronization, media stream, physical effect estimation, anchor identification, content analysis, real-world effect reproduction, XR, user experience, telecommunications networkThe patent introduces a system that acquires a media stream and identifies an 'anchor'—an element with a physical impact on the virtual scene. It then calculates the type and magnitude of this impact and commands a corresponding real-world actuator to reproduce the effect at a user's location. The innovation enhances extended reality experiences by synchronizing physical effects with the media stream in real-time.
2022-04-26 / Extended reality anchor caching based on viewport prediction / US11315326B2 pdfextended reality (XR), viewport prediction, anchor caching, digital object rendering, augmented reality (AR), virtual reality (VR), spatial computing, immersive technology, real-world overlay, future viewport determinationThis invention predicts a mobile device's future viewport in a real-world environment and pre-caches the digital objects associated with extended reality anchors located within a threshold distance of this viewport. It involves removing anchors that fall outside of the predicted area to optimize the device's processing resources. When an anchor in the remaining subset contains a digital object not yet on the device, the system downloads it from another device.
2021-11-23 / System for collaboration of augmented reality events / US11182966B2 pdfaugmented reality, AR overlay, primary user, user experience level, collaboration, system, modify, machine-readable medium, instructions, processing systemThe patent describes a system that determines the experience levels of multiple users in an AR event to identify a primary user, allows the primary user to issue commands to change the AR overlay for others, and modifies the AR overlay based on these instructions, enhancing collaborative augmented reality experiences.
2022-07-12 / Pre-fetching of information to facilitate channel switching / US11388471B2 pdfstreaming, pre-fetching, user behavior prediction, channel prioritization, media rendering, data type selection, network resources, switching latency, DRM license, media playerThe technical novelty involves selectively pre-fetching data for multiple streaming channels based on user behavior predictions, prioritizing channel data fetching based on likelihood of user request to decrease switching time between streams, and selectively pre-fetching different types of data (first and second type) for higher priority channels while fetching only one data type for lower priority channels.
2022-03-29 / Automatic generation of augmented reality media / US11290766B2 pdfaugmented reality, AR, live footage, sensor data, content analysis, analytical statistic, content filtering, computer-generated content, immersive display, context-awareA method for creating augmented reality media utilizes live footage, sensor data from event location, and historical event data to extract statistics for content analysis. It filters data to identify content of interest which is then presented as computer-generated content, overlaid on live footage in an immersive display. This synchronization results in an enhanced AR experience that augments the real-world perception with relevant, context-aware virtual objects.
2021-03-23 / Automatic generation of augmented reality media / US10958959B1 pdfaugmented reality, AR, live footage, sensor data, content analysis, analytical statistics, target filtering, immersive display, computer-generated content, real-timeThe patent details a method for creating augmented reality (AR) media by analyzing live footage and sensor data from an event, using this data to extract analytics and identify relevant content, and then generating synchronized computer-generated content to augment the live footage for immersive display. The system automatically selects content of interest by filtering through both current and historical data related to an event, allowing for real-time augmented reality experiences. This technology enables users to experience enhanced views of live events with added contextual information or virtual elements overlaid on the real-world footage.
2021-09-07 / Systems and methods for administrating suggested merchandising arrangements / US11113350B2 pdfnatural language processing, product description, personalization, user behavior history, language history, dynamic updating, online catalogs, customized shopping, NLP system, user equipmentThis method personalizes product descriptions for individual users by using natural language processing (NLP) to tailor content according to the users' language and behavior history. It dynamically updates product information presented to the user based on their responsive behavior, ensuring a customized shopping experience. The system integrates product attribute analysis with end user information to modify and present product descriptions on user equipment.
2023-03-21 / Method for content synchronization and replacement / US11609930B2 pdfcontent integration, user library, alternate content, cue points, synchronization, content replacement, playback, mash-ups, processing system, content recommendationThe innovation introduces a method for a processing system to integrate user library content into primary content at specified cue points. It allows for automatic replacement of portions of the content with recommended alternate content during playback. The system also facilitates synchronization and presentation of user-created alternate content with the original content stream.
2022-01-25 / Method for content synchronization and replacement / US11232129B2 pdfcontent playback, alternate content, processing system, content synchronization, user library, cue points, recommendations, content integration, mash-ups, user-generated contentThe technical novelty involves a processing system that allows for the integration and replacement of portions of primary content with alternate content at specific cue points based on user library matches and recommendations. This system provides a customized content playback experience by seamlessly synchronizing alternate content tracks during the playback of the primary content. It addresses the need for easy integration of user-generated mash-ups or additional content tracks with existing content sources.
2021-11-23 / Extended reality markers for enhancing social engagement / US11182965B2 pdfextended reality, XR markers, social engagement, virtual world, real world, information display, context-aware, environment modification, user interaction, automated systemThe patent describes a system for enhancing social engagement in XR environments by generating markers that convey information about other users. These markers are inferred and displayed within the combined real and virtual worlds to provide context-aware social cues to the participants. The process is automated by a computing system capable of modifying the XR environment to include these informative markers.
2023-05-02 / Method for embedding and executing audio semantics / US11640829B2 pdfsemantic metadata, audio adaptation, playback environments, content rendering, embedded semantics, processing system, adaptive audio parameters, semantic expression, consistent playback, dynamic audio contentThe patent introduces a device that dynamically adapts audio content to maintain the intended semantic expression across different playback environments by utilizing semantic metadata embedded within the content. The device includes a processing system that determines parameters for audio adaptation, ensuring that the audio content, as rendered, accurately conveys the embedded semantic intents irrespective of the playback setting. The system adapts the audio according to these parameters to provide a consistent and authentic semantic experience.
2021-07-06 / Method for embedding and executing audio semantics / US11056127B2 pdfsemantic metadata, audio adaptation, content rendering, playback environments, processing system, audio content, parameters, device, semantic expression, adaptive audioThe patent describes a device capable of adapting audio content based on parameters determined by semantic metadata embedded within the content, ensuring the content is rendered to authentically represent the intended semantics across diverse playback environments. The device includes a processing system that determines these parameters for content adaptation, then adapts and renders the audio accordingly. This technology overcomes the limitation of static content injections that may not accurately convey semantic expressions in different contexts.
2021-07-20 / Space characterization using electromagnetic fields / US11070950B2 pdfspace characterization, EM field radios, object detection, room configuration, electromagnetic spectrum, indoor mapping, navigation, wireless communication, automatic mapping, vehicle guidanceThe patent describes a system for space characterization using EM field radios to detect both static and moving objects and perform room configuration without the need for manual annotation or visual markers. This system collaborates across different EM spectra, enabling a more dynamic and automatic mapping process. The technology finds applications in navigation systems for vehicles within indoor environments.
2020-07-07 / Space characterization using electromagnetic fields / US10708718B1 pdfspace characterization, electromagnetic field radios, object detection, indoor mapping, EM spectrum, collaborative devices, environment sensing, static objects, moving objects, dynamic room configurationThe invention provides a method for space characterization using electromagnetic (EM) field radios to detect both static and moving objects within an environment without necessitating manual annotations or visual markers. This system utilizes a collaborative approach across different EM spectra, enhancing indoor mapping and object placement intelligence. The technique significantly reduces the need for initial human setup and enables a more dynamic understanding of the contents within a space.
2023-05-30 / Multi-agent input coordination / US11664032B2 pdfmulti-agent coordination, audio request identification, location context, secondary trait analysis, smart devices, listening agents, smart speaker, voice command processing, acoustic collaboration, virtual assistantThe invention improves multi-agent input coordination for acoustic collaboration by utilizing multiple listening agents within smart devices to enhance request identification and user intent understanding. A smart speaker's processor identifies which agent should process an audio request based on techniques like location context and secondary trait analysis. The system can handle simultaneous or differing audio requests from various requesters through a network of agents.
2021-11-09 / Multi-agent input coordination / US11170783B2 pdfmulti-agent, smart devices, audio requests, virtual assistants, location context, secondary trait analysis, speech processing, user intent, request identification, smart environmentsThe patent describes a system for improving request identification and user intent understanding in multi-agent smart environments by processing audio requests through multiple listening agents using location context and secondary trait analysis. This allows the system to determine the most appropriate agent for response, enhancing user experience with virtual assistants like Google Assistant™, Apple's Siri™, Amazon's Alexa™, and Microsoft's Cortana™. The technology addresses challenges with simultaneous or differing audio requests from multiple users across various devices.
2022-12-13 / Communications network security for handling proxy voice calls / US11528358B2 pdfcommunications network security, proxy voice calls, voicebot, user equipment, call handling system, verification, call request suspension, confirmation request, network attack prevention, legitimacy confirmationThe disclosed technology addresses security for proxy voice calls in communications networks by intercepting call requests from voicebots, suspending their routing, and requesting confirmation from the called devices. This system mitigates unwanted calls and potential network attacks by confirming the legitimacy of calls generated by voicebots before they reach the intended recipients. The confirmation process involves identifying both the voicebot and the associated user equipment to the target device before proceeding with the call.
2020-11-24 / Communications network security for handling proxy voice calls / US10848619B2 pdfcommunications network, voicebot, proxy voice calls, call handling system, call request interception, user equipment, voicebot confirmation request, call authorization, network security, call managementThe patent describes a call handling system that intercepts voicebot-generated call requests to a target device, suspends routing of such calls, and generates a confirmation request to validate the voicebot and user equipment. The target device thus has the option to authorize or deny the call based on the received confirmation request, enhancing security and reducing unwanted or nefarious calls. This system aims to mitigate issues of call obfuscation and increase the convenience for users managing calls from both known and unknown callers.
2021-10-19 / Selecting spectator viewpoints in volumetric video presentations of live events / US11153492B2 pdfvolumetric video, spectator viewpoints, live events, video streams, target tracking, future position prediction, video compositing, immersive experience, camera feeds, display alertThe patent describes a method for enhancing live event viewing by dynamically selecting spectator viewpoints using volumetric video, which is produced by compositing multiple camera feeds to follow a target and predict its future position, alerting a display device to shift perspective accordingly.
2022-10-25 / Service for targeted crowd sourced audio for virtual interaction / US11481181B2 pdfaudio generation, IoT, personalized audio, location-based, user preference, sound shaping, augmented reality, virtual reality, sound zones, interactive visualizationThis audio generation system integrates with IoT devices to personalize audio experiences in a targeted area by dynamically controlling audio output based on listener location and preferences. It shapes the audio landscape using envelope control and offers individualized experiences, potentially enhanced by augmented or virtual reality visualizations. Additionally, it features zone-specific sound management and an interactive interface for users to intuitively visualize and modify the sonic environment.
2021-07-06 / Method and system for stake-based event management with ledgers / US11055412B2 pdftrust-less, autonomous nodes, event information, ledger, reputation, bookmaker module, rewards, penalties, decentralized system, validationThe patent describes a decentralized system for the exchange of event information, where a network of autonomous nodes report on conditions, a ledger system maintains records of these events and the reputations of nodes, and a bookmaker module assesses and enforces rewards and penalties based on event validation or invalidation by subsequent reports from nodes.
2022-12-20 / Method to detect and counteract suspicious activity in an application environment / US11532207B2 pdfgaming fraud detection, peripheral device input, user identity, device identification, stimuli threshold comparison, validation request, proctoring, impersonation detection, remote gaming security, player authenticationThe patent describes a system for detecting fraud in gaming by comparing inputs from peripherals with a threshold value linked to a user's identity, device identification, and the stimuli presented in the game. If the input exceeds the threshold, indicating potential fraud, a validation request is sent to the user's device. The technology addresses the challenge of proctoring players and detecting unauthorized stand-ins or impersonation in both large venue and remote gaming environments.
2021-05-25 / Method to detect and counteract suspicious activity in an application environment / US11017631B2 pdfgaming security, fraud detection, collusion prevention, peripheral input, user identification, threshold comparison, validation request, player authentication, remote monitoring, anti-cheatThe patent provides a method for detecting fraud and collusion in gaming by comparing a peripheral device's input with a threshold that considers user identification, peripheral identification, and game stimuli. When inputs exceed the threshold, indicating potential fraud, a validation request is sent to the suspected user's device. This technology aims to enhance gaming security in both large venue and remote player scenarios by improving unauthorized player detection.
2021-07-13 / Synchronization of environments during extended reality experiences / US11062678B2 pdfextended reality, XR, sensors, real-world environment, virtual environment, immersive experiences, data interpolation, remote action, environment modification, extended reality deviceThis patent outlines a method for enhancing immersive experiences in extended reality (XR) by adjusting a real-world environment using data collected from sensors, which then drives modifications via signals sent to devices in proximity to the user. Data collected from the real environment is used to interpolate with a virtual environment and craft a coherent extended reality experience. Actions are triggered remotely in the real-world environment to align with the virtual experience, creating a more seamless integration between the two.
2022-05-17 / System for trend discovery and curation from content metadata and context / US11336743B2 pdfmedia content, metadata, consumer context, consumption patterns, trends analysis, user profiles, personalized recommendations, network subscribers, content discovery, targeted communicationsThe patent describes a method for personalizing media content recommendations by obtaining metadata from consumed media, establishing individual user consumption contexts and patterns, then aggregating these to discern overall trends. It correlates these trends with user profiles, including current activities, to generate targeted recommendations for novel media content. The recommendations are then communicated to the respective network subscribers for a customized content discovery experience.
2021-04-13 / System for trend discovery and curation from content metadata and context / US10979528B2 pdfmedia consumption, metadata, consumer context, consumption patterns, trend analysis, personalized recommendations, network subscribers, content aggregation, profile correlation, recommendation communicationThe invention provides a method for analyzing a network subscriber's media consumption using metadata and consumer context to establish individual consumption patterns, aggregating these patterns to discern trends, and leveraging this data to generate personalized media content recommendations, which are then communicated to the subscriber.
2021-09-28 / Methods, devices and systems for adjusting presentation of portions of video content on multiple displays based on viewer reaction / US11134288B2 pdfvideo content analysis, viewer reaction threshold, content sub-portion isolation, adaptive streaming, viewer engagement, content redistribution, individualized displays, audience interaction, real-time adaptation, engagement optimizationThis patent describes a method for analyzing viewers' reactions to video content and identifying specific content portions that trigger significant reactions. Upon determination that a viewers' reaction meets a specific threshold, the identified content sub-portion is isolated and then streamed separately to enhance engagement. This adaptive content delivery system optimizes viewer experience by focusing on and redistributing content elements with the highest impact.
2023-05-30 / Selecting viewpoints for rendering in volumetric video presentations / US11663725B2 pdfvolumetric video, video streams, trajectory, target tracking, viewpoints, composite video, viewer feedback, rendering adjustment, automated visual analysis, interactive viewingThe invention involves a method for rendering volumetric video by analyzing multiple video streams to identify and track a target of interest from various viewpoints, computing its trajectory, and creating a composite video that follows the target. Viewer feedback on the volumetric video traversal is used to adjust the rendering to improve the viewing experience. This process enhances engagement by allowing viewers to observe the scene from any chosen angle and providing a customizable viewing path that follows the target within the scene.
2021-07-27 / Selecting viewpoints for rendering in volumetric video presentations / US11074697B2 pdfvolumetric video, video streams, visual analysis, trajectory, target identification, scene rendering, viewer feedback, video compositing, multiple viewpoints, interactive videoThe technical novelty involves a method for creating a volumetric video traversal by automatically identifying a target from multiple video streams using visual analysis, determining the target's trajectory from different viewpoints, and rendering a composite video that follows the target, which then can be dynamically adjusted based on viewer feedback.
2023-06-06 / Validating objects in volumetric video presentations / US11670099B2 pdfvolumetric video, validation, objects, user-generated content, metadata, authenticity, reliability score, manipulation, filtering, integrityThe patented method involves a validation process for objects in volumetric videos by comparing these objects with user-generated content using associated metadata to confirm the authenticity of the objects. A reliability score is calculated for the objects based on the comparison, which determines whether they remain unfiltered or are filtered out of the scene if suspected of manipulation. This process aims to ensure the integrity of the volumetric video content and enhance viewer experience by maintaining authenticity.
2021-04-06 / Validating objects in volumetric video presentations / US10970519B2 pdfvolumetric video, validation, user-generated content, metadata matching, object comparison, authenticity, manipulation detection, score assignment, video filtering, enhanced viewing experienceThe patent describes a method for validating objects in volumetric video by comparing them with user-generated content that depicts the same scene. Objects are validated by matching metadata, comparing objects in both types of content, assigning a score based on similarities, and filtering out objects that are likely manipulated below a certain threshold score. This approach is used to maintain authenticity and enhance viewer experience by ensuring the reliability of objects within volumetric videos.
2022-01-18 / System for content curation with user context and content leverage / US11228806B2 pdfmedia content delivery, leverage points, user preferences, user context, content customization, personalization, media content items, content selection, custom content item, user equipmentThe patent describes a method for personalized media content delivery by extracting leverage points from media content items, utilizing user information like preferences and context to select and tailor content, and presenting the customized content to the user's equipment.
2020-08-04 / System for content curation with user context and content leverage / US10735811B2 pdfcustom content creation, media content extraction, leverage points, user preferences, user context, content selection, content refinement, personalization, media presentation, user equipmentThe method involves creating customized content by extracting leverage points from media content items, using user preferences and context to select and refine content, and presenting the tailored content to the user's equipment.
2022-10-11 / Automatic selection of viewpoint characteristics and trajectories in volumetric video presentations / US11470297B2 pdfvolumetric video, trajectory, viewpoint characteristics, video streams, rendering, compositing, immersive, interaction, customizable viewing experience, endpoint deviceThe patent details a method to automatically determine a viewer's trajectory in volumetric video presentations that involves receiving multiple video streams of a scene from different angles, identifying desired viewpoint characteristics, and calculating a trajectory that aligns with these characteristics. The method allows for rendering a personalized volumetric traversal path by compositing the video streams, enhancing viewer interaction and immersion. This technology aims to provide a more customizable and dynamic viewing experience on various user endpoint devices.
2021-05-18 / Automatic selection of viewpoint characteristics and trajectories in volumetric video presentations / US11012675B2 pdfvolumetric video, automatic selection, viewpoint characteristics, video streams, trajectories, 3D rendering, degrees of freedom, interactive viewing, compositing, video presentationThe technique describes an automated method for selecting viewpoint characteristics and plotting a viewing trajectory in volumetric video, which combines multiple video streams to create a navigable 3D scene. A trajectory is determined that aligns with predefined viewpoint preferences and allows for a composite volumetric video experience consisting of distinct video streams. The rendered volumetric video allows users to interactively view the scene with six degrees of freedom on various display devices.
2022-10-25 / Automated assistant context and protocol / US11481186B2 pdfVirtual Assistant, Communication Management Component, Interaction Tracking, Context Awareness, Adaptive Characteristics, User Engagement, Communication Devices, Dynamic Management, Interaction Enhancement, Update InformationThe patent details a system where a Communication Management Component (CMC) can track and manage interactions between virtual assistants and communication devices, as well as with other virtual assistants, by determining the interaction context. The CMC can then modify characteristics of the virtual assistant to enhance the interaction based on the determined context. Additionally, the system can provide interaction update information to newly joined communication devices or virtual assistants, enhancing their integration into the ongoing interaction.
2021-08-10 / Volumetric video creation from user-generated content / US11089281B2 pdfvolumetric video, photogrammetry, 3D representation, 2D to 3D conversion, video processing, visual communication, feature identification, endpoint devices, two-dimensional source videos, processing systemThe invention focuses on a processing system that synthesizes a volumetric video by utilizing photogrammetric principles to combine two-dimensional videos from different endpoint devices, identifying and using common features across these videos to create a three-dimensional representation. This system is designed to improve visual communication by transforming standard 2D video footages into a composite 3D experience. The processing technique ensures that shared features in the input videos are accurately merged, resulting in a seamless volumetric output.
2022-08-30 / Opportunistic volumetric video editing / US11431953B2 pdfvolumetric video, object detection, user focus, view obstruction, video manipulation, presentation, computing device, visual communication, video enhancement, user experienceThe described processing system detects objects in a volumetric video that are the focus of a user and identifies when the user's view is obstructed, subsequently performing manipulations on the video presentation to reduce or eliminate view obstructions.
2021-05-11 / Opportunistic volumetric video editing / US11006091B2 pdfvolumetric video, view obstruction, manipulating presentation, visual communication, processing system, dynamic adjustment, enhanced viewing, video presentation, object detection, user experienceThe patent pertains to a processing system that uses at least one processor to detect obstructions in the user's view of an object within a volumetric video and then manipulates the video presentation to alleviate the view obstruction. This system aims to enhance the viewing experience by dynamically adjusting the presentation based on the detection of visual obstructions. The system includes methods, computer-readable media, and devices specifically programmed for manipulating the presentation of volumetric content.
2021-12-21 / Volumetric video-based augmentation with user-generated content / US11206385B2 pdfprocessing system, 2D source video, 3D object model, volumetric video, texture mapping, enhanced three-dimensional, video enhancement, immersive video experience, volumetric space, object identificationThe patent describes a processing system that enhances 2D source videos by incorporating 3D object models into volumetric videos. It details the selection of relevant volumetric videos, identification and location determination of objects within the source video, and the subsequent texture mapping of these objects onto 3D models to create modified volumetric videos integrating enhanced 3D object models. This integration allows objects from 2D videos to be represented in a 3D volumetric space, creating a more immersive video experience.
2020-09-08 / Volumetric video-based augmentation with user-generated content / US10771763B2 pdfvolumetric video, 2D source video, 3D object model, texture mapping, video enhancement, immersive representation, object identification, video modification, spatial integration, processing systemThe invention allows for the modification of volumetric video by selecting associated features from a source 2D video, identifying and locating objects within it, then enhancing these objects using a 3D model combined with the original texturing to create a more immersive and realistic volumetric video representation.
2021-04-06 / Method and apparatus for authenticating media based on tokens / US10972777B2 pdfcontent authentication, metadata, tokenization, date, time, location, database storage, user devices, origin verification, privacy considerationsThe innovation provides a method for authenticating content by attaching a token containing metadata such as the date, time, and location of content acquisition to the content item before storage in a database. This process ensures the reliability of the origin and creation details of the content without requiring complex coordination of legal permissions and technical specifications. The tokenization of content item details simplifies authentication and retrieval from databases for verification purposes.
2021-08-31 / Method and apparatus for contextual inclusion of objects in a conference / US11108991B2 pdfvideoconference, context assessment, object identification, audience engagement, presentation enhancement, virtual objects, physical objects, presenter transition, content relevance, conference systemThis patent details a system for enhancing the engagement and understanding of audience members in a videoconference by assessing the context of the conference and identifying objects relevant to the presentation material. The system then presents these physical or virtual objects seamlessly during the conference to minimize delays and maintain audience attention and comprehension. It aids in transitioning between multiple presenters, ensuring the audience can appreciate the significance and relevance of objects to the content being discussed.
2020-02-04 / Method and apparatus for contextual inclusion of objects in a conference / US10554931B1 pdfvideoconferencing, contextual assessment, object identification, presentation enhancement, virtual objects, physical objects, presenter transition, audience comprehension, communication network, network configurationThe patent introduces a system for enhancing videoconferencing by contextually assessing, identifying, and presenting relevant objects, both physical and virtual, to reduce transition lags between presenters and clarify object relevance to the audience. This system is designed to better capture the significance of objects within the presentation’s content and streamline presenter transitions toward a more coherent audience experience. The technology may include various devices and network configurations to enable this functionality.
2021-01-05 / Network-controllable physical resources for vehicular transport system safety / US10885785B2 pdfnetwork-connected vehicle, safety, animate being, registered safety need, hazard detection, physical resource adjustment, warning system, traffic management, autonomous response, computing apparatusThe patent discloses a network-based system to enhance safety by identifying when a network-connected vehicle is a potential hazard to an individual with registered safety needs, warning the vehicle, and manipulating network-controllable physical resources to mitigate the risk. The system detects the threat and autonomously initiates a protocol to alert the vehicle and take corrective action, such as changing traffic signals, to prevent an incident. A computing apparatus is programmed to execute these functions, improving interaction between network-connected vehicles and vulnerable road users.
2023-03-14 / Video-log production system / US11605402B2 pdfvideo composition, user goals, audience preferences, video clips, network-based publishing, upload, automatic, custom video, online platforms, video productionThe patent discloses a system that composes videos tailored to user-specified goals and audience preferences by selecting from a set of video clips, and then uploads the custom video to a network-based publishing platform. The system involves processes for determining user goals and audience preferences to guide the selection and composition of video clips into a final video product. After composing the video, the system can automatically upload the final product to various online platforms for viewing and sharing.
2021-01-05 / Video-log production system / US10885942B2 pdfvideo composition, user goals, audience preferences, video clips, processing system, network-based publishing, automated upload, content tailoring, personalized video, multi-processorThe technical novelty involves a system capable of composing a video by selecting from a set of video clips, considering specific user goals and audience preferences. This system determines both the user's intent for the video and the likely preferences of the intended audience using a processing system with one or more processors. Once the video is tailored to these parameters, it gets uploaded automatically to a network-based publishing platform.
2021-09-28 / Services for entity trust conveyances / US11507955B2 pdftrust management, transaction security, device interactions, Trust Management Component, trust levels, trust conveyance, entity evaluation, adaptive trust, secure transactions, dynamic updatingThe patent details a system for managing trust levels during transactions between devices, enabling entities to convey and request trust through a Trust Management Component (TMC) based on their actions, thereby facilitating or restricting interactions according to the adjusted trust levels. Actions attempted by entities are evaluated by the TMC, which modifies trust levels and guides the transaction's progression or the response behavior between devices. This trust management system adapts to the outcomes of entity actions, dynamically updating trust to ensure secure and reliable transactions.
2021-09-28 / Services for entity trust conveyances / US11132681B2 pdftrust management, online transactions, trust levels, Trust Management Component (TMC), entity evaluation, action response, trust conveyances, transaction ecosystem, communication devices, interaction managementThis patent defines a system for managing trust levels between entities in online transactions. A Trust Management Component (TMC) evaluates and indicates entities' trust levels which influence the progress of transactions and responses to actions. Entities with insufficient trust levels can request trust conveyances, adjusting their trust levels based on the outcomes of their actions within the transaction ecosystem.
2022-11-15 / System for delivery of XR ad programs / US11501341B2 pdfinteractive, extended reality, advertisement, customization, user information, equipment capabilities, environmental context, advertisement package, feature selection, presentationThe patent describes a method for creating and displaying interactive extended reality advertisements tailored based on user-specific information. This customization integrates user data, equipment capabilities, and environmental context to select appropriate features from an advertisement package. The system then presents an optimized interactive advertising experience to the user's equipment.
2021-02-23 / System for delivery of XR ad programs / US10929894B2 pdfinteractive, extended reality, personalization, advertisement, user information, customization, engagement, equipment, environment, ad packageThe invention relates to a method for creating personalized interactive extended reality advertisements based on a user's specific information, by selecting appropriate features from an ad package with multiple options. This selection is tailored to the user's equipment and environmental context to enhance engagement. The method further includes presenting the customized interactive advertisement on the user's device.
2022-11-22 / Context-based automated task performance for user contacts / US11509659B2 pdfprocessing system, trust profile, automated system, task request, access permissions, user, requesting party, task fulfillment, virtual assistant, securityThe patent describes a processing system that evaluates a requesting party's trust profile before allowing access to a first automated system to fulfill task requests on behalf of a user. The system determines access permissions based on the trust profile of the requesting party, which ensures secure and selective task processing. The system then completes the task through the allowed automated system, contingent on trust profile validation.
2022-05-24 / Identifying and addressing offensive actions in visual communication sessions / US11341775B2 pdfprocessing system, communication session, user representation, real-time, visual content, action detection, modification, predefined settings, video communication, interactiveThe patent describes a processing system configured to facilitate a communication session by detecting actions within visual content representing a user and modifying this content based on the user's predefined settings before transmitting to another user. This technology allows for real-time alterations of a user's visual representation during a video communication session. Adjustments made can range from simple cosmetic changes to complex interaction modifications, enhancing the communication experience.
2021-02-16 / Identifying and addressing offensive actions in visual communication sessions / US10922534B2 pdfprocessing system, communication session, visual content, user representation, action detection, content modification, user configuration, transmission, personalization, visual communicationThe patent describes a system for establishing communication sessions with visual content, where actions of a user's visual representation are detected and modified based on user configurations. This personalization modifies the user's actions within the visual content before transmitting it to another user's communication system. The technology involves action detection models and configurable user settings to enhance interaction within a visual communication session.
2023-05-30 / Physical object-based visual workspace configuration system / US11665309B2 pdfvisual communication, video content modification, visualization actions, object detection, camera, user-configured settings, personalized environment representation, communication session, physical object, video transmissionThe patent describes a system to enhance visual communication by modifying video content using visualization actions tied to physical objects within a user's environment, based on user-configured settings. The modification process involves object detection within video captured by a user's camera and applying visualization actions to these objects before transmitting the video to a second user. The system is configured to allow for personalized and altered representations of physical environments during communication sessions.
2022-03-29 / Physical object-based visual workspace configuration system / US11290659B2 pdfvideo content modification, communication session, user-specific configuration, visualization actions, physical object detection, altered video transmission, dynamic visual experience, personalized video calls, environment object display, processing systemThe patent details a method for modifying video content during a communication session between two users, based on user-specific configuration settings that dictate visualization actions for detected physical objects within the video content. A processing system detects objects within video footage from a user's environment, applies the determined visualization actions to the objects according to the user's settings, and transmits the altered video to the second user's communication system. The system is designed to personalize the visual experience of video calls by dynamically altering the display of environmental objects.
2020-11-03 / Physical object-based visual workspace configuration system / US10827132B2 pdfprocessing system, video content modification, communication session, physical object detection, first user, second user, visualization action, camera, user configuration, visual communication enhancementThe patent involves a processing system that modifies video content of a user's physical environment by detecting objects and applying user-specified visualization actions during a communication session with another user. The system obtains video content through a camera, applies modifications based on pre-set configurations, and transmits the altered video to the second user's communication system. The method aims to enhance visual communication by allowing personalized alterations of the visual content shared between users.
2019-12-03 / Physical object-based visual workspace configuration system / US10498973B1 pdfprocessing system, video content modification, real-time communication, user configuration settings, physical object detection, visualization actions, video feed alteration, personalized visual modifications, dynamic video processing, communication technologyThe patent details a processing system that facilitates modified visual communication sessions between users by dynamically altering video content based on user-specific configuration settings. The system detects physical objects within a user's video feed, applies predefined visualization actions to those objects, and transmits the altered content to the other user's communication system. This allows for personalized visual modifications within the context of a real-time video conversation.
2022-03-22 / Blockchain based information management / US11281796B2 pdfblockchain, sensitive information, access management, personal data, query count, transaction conditions, decremented access, data privacy, controlled dissemination, information ownershipThis invention presents a blockchain-based method for managing access to sensitive personal information by defining a transaction that incorporates data and access conditions, including a decremented query count limit. Access to such information by requesters is regulated by the satisfaction of the set conditions and tracked via the blockchain, ensuring controlled dissemination of personal data. Once the allowable number of queries is reached, as defined in the conditions, access is restricted, thereby enforcing the owner's control over their sensitive information.
2021-11-30 / Autonomous vehicle signaling system / US11188094B2 pdfvisual indicator system, autonomous vehicles, sensor data, environmental interaction, safety, signal processing, visual signals, memory system, vehicle surroundings, decision-makingThe technical novelty of the patent lies in an autonomous vehicle visual indicator system that employs sensor data to identify when certain conditions are met and accordingly provides a corresponding visual indication to the physical environment. The system is designed to enhance the safety of autonomous vehicles by accurately conveying sensor-assimilated information regarding the vehicle's surroundings through specific visual signals. It uses a combination of sensors, processing system, memory system with executable instructions, and visual indicators to improve decision-making and inform the environment of the vehicle's actions.
2023-05-09 / Controlling access to datasets described in a cryptographically signed record / US11645408B2 pdfdataset validation, cryptographically signed record, dataset exchange, access control, compliance standards, data integrity, privacy, transformation monitoring, data management systems, availability requirementsA dataset validation platform employs cryptographic signing to maintain a record of datasets, regulating access based on each dataset's availability requirements and compliance standards. The platform updates the signed record with dataset requests, transformations based on datasets, and changes to datasets' availability criteria, ensuring adherence to privacy and usage policies. Monitoring of data transformations is facilitated by generating reliable records, enforcing compliance and protecting data integrity during exchange or transformation between different data management systems.
2021-08-31 / Controlling access to datasets described in a cryptographically signed record / US11106812B2 pdfdataset validation, cryptographically signed record, data exchange environment, access control, availability requirements, privacy limitations, compliance standards, dataset transformation, data management, monitoringThe dataset validation platform monitors and controls access to datasets in a data exchange environment by updating a cryptographically signed record, reflecting actions such as access requests, dataset transformations, and changes in datasets' availability requirements. Compliance with specified availability requirements, which may include privacy limitations or other usage criteria, is enforced when granting access to datasets. Record updates also extend to track modifications occurring due to dataset transformations or changes in their compliance standards.
2021-07-20 / Persona/individual based actions based on community specific trigger / US11070430B2 pdfautomation, networked communities, distributed job, trigger detection, automated action sets, community management, IoT, remote execution, member devices, orchestrationThis patent describes a method for automating actions within networked communities by selecting member devices associated with a distributed job, detecting triggers related to that job, and then spawning automated action sets in response to these triggers, which can be executed either by the member devices themselves or a remotely located device.
2023-02-14 / Task delegation and cooperation for automated assistants / US11579923B2 pdftask delegation, automated assistants, centralized support center, sub-task formulation, capability determination, bots, task optimization, cooperative system, complex task execution, digital assistant specializationThe patent introduces a method for dividing tasks requested by an individual into sub-tasks and intelligently delegating these sub-tasks to an array of specialized automated assistants based on each assistant's capabilities. A centralized support center assesses and distributes these tasks to optimize performance and ensure that each function is handled by the most efficient digital assistant available. This process facilitates a cooperative system among various automated assistants to achieve complex task executions that may be beyond the capability of a single assistant.
2022-05-03 / Task delegation and cooperation for automated assistants / US11321119B2 pdftask delegation, automated assistants, centralized support, sub-tasks, collaboration, capability-based assignment, task efficiency, digital assistants, complex tasks, cooperationThe patent describes a system for task delegation among automated digital assistants coordinated by a centralized support center capable of breaking down a task into sub-tasks and assigning them to different assistants based on their capabilities. The innovation allows for efficient collaboration between multiple automated assistants to complete complex tasks on behalf of individuals more rapidly and effectively than a single assistant. This centralized approach leverages the strengths of various automated assistants to improve task execution speed and quality.
2020-10-13 / Task delegation and cooperation for automated assistants / US10802872B2 pdftask delegation, automated assistants, centralized support, sub-tasks, bot capabilities, cooperative completion, artificial intelligence, multi-bot coordination, task subdivision, automationThis patent introduces a system that enables centralized task formulation and delegation among multiple automated assistants, allowing complex tasks to be subdivided and assigned to specific bots based on their capabilities, thus facilitating cooperative task completion.
2022-05-10 / Edge sharing orchestration system / US11330050B2 pdfnetwork device, processor, memory, edge share orchestrator, edge devices, computing power, functionality augmentation, service performance, resource meshing, customer device enhancementThe patent describes a network device with a processor and memory that initiates an edge share orchestrator to identify edge devices and enhance a customer device's capabilities. It compensates for a customer device's lack of computing power or functionality by leveraging additional resources from other edge devices. The orchestration process includes meshing additional device resources with the customer device to perform services that the customer device alone could not handle.
2020-08-11 / Edge sharing orchestration system / US10742728B2 pdfedge share orchestration, device pooling, capability integration, unused computing resources, service augmentation, local network optimization, opportunistic capability listener, edge share balancer, dynamic resource allocation, user experience enhancementThe patent describes an edge share orchestration system that connects and pools together various customer and additional devices based on their capabilities to create new or enhanced services. By utilizing idle compute power and functionalities of underutilized devices, the system aims to optimize resource usage within a local network. The orchestration system ensures that each device contributes to the user experience by dynamically integrating and leveraging their combined capabilities.
2023-05-16 / System for active-focus prediction in 360 video / US11651546B2 pdffield of view prediction, viewer line of sight, points of interest scoring, media content presentation, bandwidth optimization, 360-degree video, viewing habits, content streaming, bandwidth conservation, media consumption efficiencyThe patent describes a system for enhancing viewer experience and reducing bandwidth usage by predicting a viewer's field of view in media content, such as 360-degree video, and updating the scoring of points of interest based on the tracked line of sight. The predicted field of view is refined with updated scoring, optimizing what content is streamed or presented to the viewer. This method allows for bandwidth conservation by focusing on the parts of the video that the viewer is most likely to watch.
2021-08-24 / System for active-focus prediction in 360 video / US11100697B2 pdffield of view prediction, viewer line of sight, media content, point of interest scoring, bandwidth optimization, 360-degree video, real-time updating, network resources, viewing optimization, content presentationThe patent outlines a method for reducing bandwidth usage in presenting media content, such as 360-degree video, by predicting a viewer's field of view and updating scoring for points of interest based on the viewer's line of sight. This approach allows for dynamically adjusting the presentation of media content to the viewer's most likely field of interest, thus optimizing bandwidth. The system refines the field of view prediction and point of interest scoring in real-time, enhancing the viewer's experience while conserving network resources.
2020-09-22 / System for active-focus prediction in 360 video / US10783701B2 pdf360-degree video, bandwidth reduction, field of view prediction, eye movement tracking, content delivery efficiency, video streaming, viewer interest analysis, line of sight monitoring, adaptive streaming, video content analysisThe innovation reduces bandwidth consumption in 360-degree video streaming by predicting and sending only the viewer's predicted field of view to their device, which is less than 180 degrees, based on their interest and eye movement tracking. The technology continually updates this prediction as it monitors the viewer's line of sight, leading to a more efficient content delivery. This approach addresses the bandwidth inefficiencies of streaming full 360-degree videos when users only view a portion of the available content.
2019-11-19 / System for active-focus prediction in 360 video / US10482653B1 pdfbandwidth optimization, media content analysis, user field of view, interest point scoring, predictive modeling, line of sight monitoring, 360 degree video, user information, dynamic updating, media content deliveryThe patent introduces a method for optimizing bandwidth usage in media content delivery by scoring points of interest within the media, predicting a user's field of view based on user information and a previously analyzed interest point, and dynamically updating this prediction by monitoring the user's actual line of sight.
2022-05-03 / Data parking within offline community system / US11323896B2 pdfemergency communication, offline device management, data aggregation, community of devices, mobile aggregation device, artificial intelligence, virtual network function, network device, event assessment, recovery operationsThis patent outlines a method for managing emergency communication scenarios where devices go offline by defining a community of devices that includes offline devices and utilizing a mobile aggregation device to maintain data/message flow. By invoking an emergency mode, the system allows for the aggregation of communication payloads from the community of devices, while incorporating artificial intelligence as a virtual network function to assess and respond to the events. The AI leverages the aggregated data from the mobile aggregation device to evaluate the situation, enhancing recovery operations.
2020-10-06 / Data parking within offline community system / US10798592B2 pdfdata parking, offline community system, mobile aggregation device, disaster recovery, intermittent connectivity, data aggregation, network resumption, emergency communication, data distribution, network devicesThe invention provides a system for data parking in offline scenarios, utilizing a mobile aggregation device that moves between network-connected and disconnected positions to aggregate and store data from offline devices within a community. The mobile aggregation device reestablishes communication with offline devices during events such as power outages or natural disasters and retains the data until it can return to a network-connected position. This system ensures efficient data aggregation and distribution among network components when intermittent connectivity issues occur.
2023-02-28 / Gaze-based workflow adaptation / US11593725B2 pdfworkflow adaptation, gaze-based, automation, user engagement, task monitoring, remote device, trigger actions, physical and digital tasks, gaze detection, automatic task executionThe patent describes a method that facilitates automated workflow adjustments based on a user's gaze, specifically by monitoring the user's progress within a workflow, detecting when the gaze meets certain criteria, and triggering predefined actions in a remote device. This gaze-based interaction enhances efficiency in workflows that involve a mix of physical and digital tasks. It enables automatic task execution through gaze detection, eliminating the need for manual initiation of the next steps in a workflow.
2023-06-27 / Content distribution pipeline watermarking / US11689782B2 pdfmedia content tracking, watermarking, network node, processing system, access tracking, content access, endpoint device identification, embedded watermark, digital rights management, media access notificationThe patent describes a system for tracking media content access using watermarks, where a processing system receives a watermark from a network node embedded in the media content, and a notification that includes the watermark and identification of the accessing device, in order to record the content access.
2021-10-05 / Content distribution pipeline watermarking / US11140460B2 pdfmedia content tracking, watermark embedding, network node, media access, endpoint device identification, access logging, content distribution, watermark detection, processing system, media usage monitoringThe invention provides a system for tracking access to media content by embedding a watermark at the network node level, and subsequently detecting this watermark upon media access to record such events. When the media is accessed by an endpoint device, a notification including the embedded watermark and the device's identification is sent to a processing system. This system then logs the media content access by the endpoint device, allowing for precise monitoring of content distribution and usage.
2022-01-04 / Directing user focus in 360 video consumption / US11218758B2 pdfmedia content personalization, highlight points, degrade points, user information, content modification, user engagement, consumption monitoring, customized content, attention direction, media enhancementThis patent relates to a method for personalizing media content consumption by obtaining media content, identifying user-specific points of interest, and modifying the content to highlight or degrade certain aspects based on user information. The personalized highlighted content is presented to the user to direct their attention towards designated highlight points and away from degrade points, potentially enhancing user engagement. User consumption of this customized content is further monitored for feedback or analysis.
2020-07-21 / Directing user focus in 360 video consumption / US10721510B2 pdfmedia content customization, user engagement, highlight points, degrade points, user preferences, content modification, personalization, attention direction, user interaction, enhanced consumptionThe invention pertains to a method for customizing media content presentation based on user information, by emphasizing selected points of interest (highlight points) and downplaying others (degrade points). This personalized media modification process is designed to enhance user engagement by directing attention to content areas matched to their preferences, while also mitigating the issue of users potentially missing important content. The system also tracks user interaction with the modified content for further refinement.
2021-10-26 / Automated learning of anomalies in media streams with external feed labels / US11157524B2 pdfanomaly detection, media content, event type, signature generation, metadata correlation, real-time alerting, automatic labeling, signature matching, event notification, content analysisThis invention pertains to the detection and notification of anomalies in media content, where a system detects an anomaly within media content, creates a signature for it, and correlates it with specific event types using associated metadata such as time and location. Once an anomaly signature is established, subsequent media contents are scanned for matching anomalies, and upon a match, a notification is transmitted indicating the associated event type. This allows for real-time alerting and correlation of specific types of events through analysis of media content anomalies.
2021-12-14 / Automated virtual network function modification / US11201798B2 pdfvirtual network functions, VNF, automated modification, virtualization, replicated environments, performance testing, resource management, network optimization, controlled testing, virtual machinesThe technical novelty lies in creating systems and methods for automated modification of virtual network functions (VNFs) by replicating environments to compare performance of modified VNFs against originals prior to deployment. This enables administrators to manage and optimize VNFs effectively, addressing variabilities across different environments and user needs. The approach includes measuring and testing the changes in a controlled manner, which facilitates better-informed decision-making in the management of virtualized network resources.
2021-03-23 / Dynamic derivation of and broadcast to emergency contact list / US10959082B2 pdfemergency communication, dynamic contact group derivation, anomalous behavior detection, network entity messaging, infrastructure limitation mitigation, automatic message dispatch, mobile connectivity, network overload, contact group creation, user behavior monitoringThe patent describes a system capable of automatically recognizing emergency situations or anomalous user behavior, constructing a relevant contact group, and dispatching messages to this group via a network entity. It addresses infrastructure challenges during crises by optimizing communication to community-of-interest connections despite limited mobile connectivity or network overloads. The system dynamically derives contact groups based on the context, ensuring critical communication is maintained during emergencies or upon detection of unusual behavior.
2020-07-14 / Dynamic derivation of and broadcast to emergency contact list / US10715987B2 pdfemergency communication, dynamic contact group, network congestion, confirmation message, mobile connectivity, community of interest, status update, infrastructure damage, communication system, contact managementThis patent describes a system that facilitates reaching out to predefined contact groups in the event of an emergency, when network connectivity is impaired. The system dynamically creates and sends out confirmation messages to designated contacts, circumventing typical network congestion during such situations. It intelligently manages communications to ensure timely status updates between individuals and their community of interest.
2023-05-16 / Drop-in probe that facilitates management and configuration of internet of things network connected devices / US11653201B2 pdfinternet of things, IoT, proxy device, security synchronization, security updates, network management, configuration, software updates, low-power devices, network integrityThis patent introduces a proxy device system for managing Internet of Things (IoT) network devices, specifically by performing security synchronization and facilitating security updates. The proxy device identifies IoT devices within a network, taking on delegated responsibilities for security update synchronization. The system efficiently resolves the challenge of managing software updates for low-power, intermittently inactive IoT devices, enhancing security and network integrity.
2021-08-03 / Drop-in probe that facilitates management and configuration of internet of things network connected devices / US11082837B2 pdfIoT, proxy device, network security, security updates, device management, operational parameters, synchronization, communication network, identity determination, configurationThe technical novelty involves a proxy device that manages security updates for multiple Internet of Things (IoT) devices within a network by determining their identities and operational parameters and facilitating security synchronization. This device ensures both active and inactive IoT devices receive necessary security updates, mitigating the issues of unsynchronized updates and enhancing overall network security. The system is provisioned to operate within a specified communication network and assume delegated responsibility for security update synchronization.
2020-11-10 / System for management of requirements-based advertisements / US10832275B2 pdfadvertising verification, user device, ad placement, quality assessment, presentation context, experience evaluation, display monitoring, content consumption, ad effectiveness, provider criteriaThe patent provides a method for monitoring and verifying the quality and context of advertisement presentations across user devices, which includes assessing an ad's placement, size, timing, and duration on-screen, as well as analyzing surrounding content, user location, and activity during ad display. The method entails evaluating the user's advertising experience by comparing it with preset criteria from the ad provider, ensuring adherence to advertising contracts. The system verifies ad display to the provider and enhances the effectiveness of the advertisement by confirming its correct delivery and context based on predetermined parameters.
2021-03-23 / Providing logistical support for robots / US10953541B2 pdfrobotics, logistical support, centralized support center, error resolution, task performance, query processing, external analysis, proposed solution, communication, automationThe patent describes a system whereby robots communicate with a centralized support center to address errors encountered during tasks, utilizing a system which receives error reports, enables solicitation of solutions from external parties, and delivers proposed solutions back to the robots.
2021-06-22 / Network-controllable physical resources for sensory service / US11040290B2 pdfbiometric data, sensory service, group users, tailoring experience, network-controllable, physical resource, entertainment rides, dynamic adaptation, data aggregation, user experienceThe invention provides a process to customize sensory services by collecting and analyzing biometric data from a user group, then adjusting network-controllable physical resources accordingly. The system aggregates individual user biometric data to create a group profile, which is used to tailor the sensory experience. This allows for dynamic adaptation of services like entertainment rides to enhance group user experience based on aggregated biometric feedback.
2022-08-09 / Content delivery and consumption with affinity-based remixing / US11412010B2 pdfcontent customization, user reaction, sensor analysis, real-time adaptation, affinity grouping, social context, physical environment, enhanced presentation, social network sharing, affinity analysisThis invention relates to customizing content for a user by analyzing their physical and social contexts and their reactions to content, adapting the content accordingly, and then sharing the modified content with others sharing similar affinities. The system utilizes sensors to gather user reactions, adjusts the content in real-time based on this data, and identifies an affinity group for distributing the enhanced content through a social network. Enhancement of content is determined by increased user affinity upon successive interactions, and grouping is performed by analyzing reactions from a broader user base.
2021-10-26 / Content delivery and consumption with affinity-based remixing / US11159585B2 pdfcontent modification, user affinity, context-aware, reaction analysis, sensor data, social network, real-time adaptation, affinity group, dynamic content distribution, physical and social environmentThis method modifies content during presentation based on a user's physical and social context, by analyzing reactions obtained from sensors to assess affinity. Content is then modified accordingly and shared with other users on a social network if affinity is enhanced, to establish an affinity group among users with similar reactions. The system dynamically tailors and distributes content, utilizing both initial user reaction and subsequent affinity responses from a wider social network.
2022-08-09 / Video curation service for personal streaming / US11410702B2 pdfvideo analysis, machine learning, object detection, video clipping, automated editing, real-time processing, notifications, user input, keyword identification, content sharingThe patent discloses a video analysis system using a machine learning algorithm to create video clips by identifying objects in a video as they are being recorded, using a keyword received as a user input. The system determines start and stop points for these clips, automates the editing process, and sends notifications upon creation of the clips, thereby reducing the need for manual review. The system enhances searchability and sharing of relevant video content on social networks.
2020-11-03 / Video curation service for personal streaming / US10825481B2 pdfautomated video editing, machine learning, object detection, real-time, video clips, user input, keyword identification, notification system, content sharing, video recordingThe invention pertains to an automated video editing device that utilizes a machine learning algorithm to create video clips with defined start and stop points based on real-time object detection during video recording. The system receives specific user input keywords to identify objects for monitoring and enables automated generation of relevant video clips without user engagement in editing. Upon creating a video clip, the system automatically notifies the user, simplifying content sharing and review processes.
2023-01-10 / Context-based object location via augmented reality device / US11551444B2 pdfaugmented reality, visual feed, user context, object indication, personalized enhancement, processing system, visual assistance, context detection, AR display, enhanced visualizationThe patent describes a system and method for augmenting a visual feed based on detected user context to enhance object indication through an augmented reality display. The improvement lies in detecting context from the feed and using this information to locate and highlight objects in a manner personalized to the user's situation. This technology integrates with processing systems and augmented reality to provide a tailored visual assistance experience.
2020-05-12 / Context-based object location via augmented reality device / US10650239B2 pdfaugmented reality, visual feed, object detection, user context, enhanced indication, real-time, AR display, processing system, environment interaction, visual enhancementThe patent introduces a system that utilizes a processing system with one or more processors to detect a user's context via a visual feed, then locate and provide an augmented reality (AR) enhanced indication of an object relevant to the user's context. This enhanced indication aims to assist users in identifying and interacting with objects in their environment. The system integrates user context with augmented reality technology to improve object detection and visualization in real-time scenarios.
2022-07-05 / Event-based community creation for data sharing platform / US11379615B2 pdfdata sharing community, event-triggered, permission levels, real-time, data sources, data consumers, network-based, data collection, access control, event typeThe patent discloses a system for dynamically creating a data sharing community triggered by an event, involving a multitude of data sources and consumers with varying permission levels. This system automatically collects and distributes event-related data to authorized parties based on predefined permissions. The novelty lies in the real-time, event-driven formation of the data sharing ecosystem with controlled access.
2021-12-07 / Network routing of media streams based upon semantic contents / US11196669B2 pdfsemantic content detection, media streams, machine learning, destination device selection, routing, media source device, content-aware distribution, streaming media networks, computer-readable media, processing systemA system utilizes machine learning models to detect semantic content within media streams and subsequently routes these streams to appropriate destination devices based on the identified content. The process includes detection of semantic content, selection of destination devices tailored to the detected content, and the forwarding of media streams to these selected devices. This enables dynamic and content-aware distribution of media streams within streaming media distribution networks.
2022-05-24 / Methods, systems and devices for selecting advertisements based on media profiles and advertisement profiles / US11341540B2 pdfmedia profile, media metrics, advertising, advertisement selection, media content, ad metrics, advertisement profiling, targeted presentation, playback device, advertising pricingThe invention relates to a method for optimizing advertisement selection for media content by creating media profiles with media metrics for different portions of the content, determining pricing for advertising, and matching advertisements with complementary ad metrics and suitable pricing to the media profiles for targeted presentation.
2020-11-17 / Network aggregation of streaming data interactions from distinct user interfaces / US10841647B2 pdfnetwork aggregation, streaming data, multimedia content, content segmentation, boundary time marker, segment overlay field, metrics collection, video splicing, user interface, data managementThe patent describes a system for aggregating multimedia content into segmented packages that are configured with overlay fields to collect metrics tied to boundary time markers, facilitating the management and analysis of streaming data interactions from multiple user interfaces. The system employs a network server to splice video content into segments, each associated with a time-specific overlay field enabling client-generated metrics collection. This allows for content distribution to user equipment with enhanced interactivity and data aggregation capabilities.
2020-11-17 / Audiovisual content screening for locked application programming interfaces / US10841652B2 pdfcontent screening, locked API, media application, audiovisual content, scraping, user equipment, streaming, display, accessing content, content analysisA system provides content screening for media applications with locked APIs by detecting content streams on user equipment, accessing and scraping audiovisual content presented on the device's display, circumventing the need to access restricted APIs, and creating a dataset of the content for analysis or control during a specific time period.
2020-06-23 / Audiovisual content screening for locked application programming interfaces / US10694250B2 pdfstreaming content, locked API, media application, content scraping, user equipment, display capturing, ratings filter, audiovisual analysis, content control, programmatic accessThe patent describes a system and method for screening and filtering streaming media content by capturing and analyzing audiovisual content directly from the display of a user equipment, bypassing locked application programming interfaces of media applications. This allows for content control beyond static ratings limits by scraping displayed content for a time period to create a processed version that can be evaluated for appropriateness. The technology enables oversight of streaming content without needing access to or modification of the media application's secure API.
2021-09-07 / Content curation for course generation / US11113472B2 pdfeducational courses, content segments, core topics, similarity threshold, user request, course assembly, content sources, subject analysis, online education, processing systemA system to assemble educational courses automatically identifies and evaluates content from various sources on a subject, determining core topics based on frequency and similarity measures. It constructs a course by including content segments that cover these core topics upon a user's request. The assembled course is then presented to the user's device for educational purposes.
2022-05-31 / Method and apparatus for content curation / US11350148B2 pdfcontent curation, user profile, media consumption, user feedback, user context, media content, personalized media channel, content service, presentation device, media exclusionThe patent relates to a content curation system which updates a user's profile based on their media consumption and feedback, determines the user's current context through their device and activities, and then modifies available media content to present a personalized media channel that excludes certain types based on the user context.
2020-12-01 / Method and apparatus for content curation / US10856028B2 pdfmedia content curation, user profile updating, user context determination, personalized media channel, content discovery, adaptive content selection, media consumption feedback, intelligent content filtering, user activity recognition, telecommunicationsThis patent describes a system for curating media content by updating a user profile based on the user's media consumption and feedback, determining the user's current context from their device, and adjusting available media content accordingly to create a personalized media channel. The system eliminates content types from the selection based on determined user context, improving the content discovery process. The method adapts to various devices, locations, and user activities to cater to changing content preferences.
2020-02-18 / Collaborative augmented reality system / US10565764B2 pdfaugmented reality, AR profiles, annotation content, collaborative system, computational processing, similarity threshold, user devices, real-world objects, local database, rendering instructionsThe invention provides a collaborative augmented reality (AR) system that uses a local database to match real-world objects with AR profiles, enabling the rendering of related annotation content on user devices. By using a second user's AR profile for an object and a similarity threshold, the system delivers relevant AR annotations to enhance the user's real-world experience. This process addresses the computational intensity and developer challenges associated with standalone AR applications through collaborative user and object data sharing.
2020-10-13 / Collaborative visual enhancement devices / US10803610B2 pdfcollaborative object detection, image processing, feature capture, device orientation, bearing determination, object location, telecommunication, visual assistance, environmental searching, device networkingThe patent describes a collaborative object detection system where a processing system detects an object within a captured image, captures the object's features, and transmits this data along with the device's location and bearing to a second device. The system helps orient user-operated devices towards the detected object by utilizing the transmitted object features, the second device's location, and its bearing to the object. This allows for enhanced collaborative visual searching using connected devices to improve object finding in various environments.
2022-06-14 / Predictive search with context filtering / US11360967B2 pdfbiometric condition, threshold, triggering object, camera, context filters, personal parameters, relevance topic, search terms, user interface, machine learningThe method detects when a user's biometric condition exceeds a set threshold, associates it with a triggering object identified by a camera, and applies context filters involving this object and user-specific parameters to generate relevant search term recommendations. These filters analyze the triggering event in conjunction with personal parameters to determine the contextual relevance of topics. The system then presents tailored search terms to the user for improved query outcomes based on these contextual insights.
2020-03-24 / Predictive search with context filtering / US10599640B2 pdfbiometric condition, threshold, triggering object, camera, personal parameter, context filter, relevance topic, search terms, user interface, tailored recommendationsThis method involves detecting a user's biometric condition that exceeds a certain threshold and associating it with a triggering object captured by a camera, then using the object and personal parameters to apply context filters that determine a relevance topic. Based on the relevance topics, a set of tailored search terms are generated. These search terms are then displayed on a user interface to assist the user in finding pertinent information more effectively.
2021-10-26 / Deviations from workflows associated with physical items / US11157845B2 pdfworkflow deviation, image capture, physical item, process monitoring, notification system, corrective action, assembly, repair, configuration, methodThe patent introduces a method and system that uses image capture to monitor tasks performed on a physical item according to a predefined workflow and utilizes a processing system to identify workflow deviations. Upon detection, the system generates a notification of the deviation, thereby prompting a corrective action response. The system's monitoring enhances the integrity and accuracy of physical item assembly, repair, or configuration processes.
2020-10-20 / Methods, systems and algorithms for providing anonymization / US10810324B2 pdfanonymization service, privacy as a service, personal preferences, communication channels, public disclosure, commercial access, user interaction, web portal, data control, anonymous transactionsThe technical novelty lies in providing a service-oriented anonymization framework that allows individuals to control the public disclosure and sale of personal preferences, such as likes and dislikes, across various communication channels including emails, web activity, browser history, telephone calls, and physical mail. The framework encompasses methods and systems for offering privacy as a commercially available service, with the flexibility to permit or restrict public and commercial access to personal information. The service integrates with various communication systems and utilizes a web portal for user interaction to ensure anonymity in these transactions.
2021-01-05 / Data and context based role membership system / US10887315B2 pdfrole management, dynamic assignment, discovery module, event response, temporary roles, credentials, community network, automatic promotion, situational authority, adaptive systemThe invention introduces an adaptive role management system that automatically designates temporary roles to individuals in response to specific events through a discovery module that evaluates their credentials, functions, capabilities, and historical response data within a community network. This method expedites access to necessary data and authoritative functions during critical situations by dynamically promoting a responder, eliminating delays typically associated with manual approvals in traditional systems. The system is uniquely capable of opportunistically and temporarily assigning functional attributes to individuals based on situational necessity, to enhance responsiveness in a dynamic network environment.
2019-08-06 / Wireless energy transfer in a far-field environment / US10374470B1 pdfwireless power transmission, antenna system, far-field, metamaterials, beam shaping, predicted positioning, mobile device charging, Fresnel zone, position tracking, wireless charging technologyThis patent discloses an improved antenna system for far-field wireless power transmission that utilizes metamaterials to focus the power beam and an antenna controller to track and aim at the mobile device's predicted position. Position, motion, direction, and height information are utilized for accurately predicting the mobile device's location to facilitate efficient power transfer in the Fresnel zone. The system enhances the feasibility of charging mobile devices wirelessly at large distances where traditional near-field methods are ineffective.
2021-11-02 / Proximity based data access restrictions / US11165771B2 pdftemporary authorization, user interaction, data set access, access record, authorization revoking, secure data modification, electronic data access, insecure location detection, processor, context-based accessThe patent details a method for providing a second user temporary access to a data set initially authorized for a first user, by detecting their interaction and generating an access record, followed by revoking access when the interaction ends. A secure procedure modifies restricted data visibility if an authorized user attempts access from an insecure location. The system employs at least one processor to manage user interactions, temporary authorization, access recording, and secure data modification based on the context of access.
2021-10-05 / Object modeling with adversarial learning / US11137761B2 pdfobject model, unmanned vehicle, physical properties, measurements, adversarial example, feature space, prediction, autonomous vehicle, model updating, computing deviceThe patent describes a method for enhancing an object model by capturing measurements with unmanned vehicles and incorporating adversarial examples. The improved model results from adding new measurements when predictions using the model fail to recognize adversarial examples. This iterative process strengthens the model's accuracy in representing the object's physical properties.
2022-04-12 / Predictive network bandwidth adjustment based on identified future event and corresponding network service / US11303581B2 pdfsoftware defined networking, bandwidth prioritization, quality of service, intelligence agent, virtual network functions, premise network, dynamic bandwidth allocation, network services, user equipment, cloud computingThe disclosed software defined networking system introduces an intelligence agent on a gateway device or server providing dynamic bandwidth prioritization and service quality control for user equipment on premise networks. It also enables instantiation of virtual network functions locally or on cloud servers to support varying network service needs. This system expands upon existing network management solutions by extending quality of service and bandwidth prioritization from mobile networks to user and premise networks.
2020-11-17 / Prioritized network based on service and user device / US10841238B2 pdfsoftware defined networking, SDN, bandwidth prioritization, premise network, intelligence agent, virtual network functions, quality of service, gateway device, cloud server, user equipmentThe invention relates to a software defined networking (SDN) system that prioritizes network service and bandwidth for devices on a customer's premise network, using an intelligence agent residing on a gateway device or a server. It dynamically adjusts bandwidth and can instantiate virtual network functions on the gateway or a cloud server to support network services for user equipment. The system extends quality of service and bandwidth management capabilities traditionally limited to mobile networks to user-centric premise networks.
2020-11-10 / Monitoring food intake / US10832590B2 pdffood intake monitoring, sensor data collection, nutrient consumption estimation, dietary recommendations, personalized nutrition, health management, eating habits, automated assistance, target nutrient goals, user-centricThis invention details a system incorporating sensors to monitor an individual's food consumption, utilizes collected data to estimate current nutrient intake, and generates personalized dietary recommendations to assist the user in achieving specific nutritional goals.
2021-11-30 / System and method of providing automated customer service with augmented reality and social media integration / US11188188B2 pdfaugmented reality, AR presentation, 3D object detection, sensor, display, processing system, tracking, enhancement, virtual reality, user interactionThe patent discloses a device that combines a display, a 3D object detecting sensor, and a processing system to create and present augmented reality (AR) enhancements of detected 3D objects to a user. The device identifies and tracks the position of a 3D object in proximity to the user, then overlays this real-time data with virtual enhancements to generate an interactive AR experience. This is delivered through a display system, potentially improving user interaction and comprehension in various applications like business services, where technicians encounter complex installation and repair scenarios.
2020-10-27 / System and method of providing automated customer service with augmented reality and social media integration / US10817133B2 pdfaugmented reality, 3D object tracking, AR presentation, sensor, processing system, user interaction, environmental input, virtual highlighting, object enhancement, real-time displayThe patent introduces an augmented reality (AR) device capable of detecting and tracking three-dimensional (3D) objects, and enhancing the representation of these objects within an AR environment for display to the user. This device leverages a sensor and processing system to identify objects in proximity to the user, create augmented presentations with enhancements, and integrate these into the user's real-time view. The system optimizes user interaction by providing augmented overlays and iconography on tracked objects to aid in tasks such as product analysis or equipment installation.
2019-12-17 / System and method of providing automated customer service with augmented reality and social media integration / US10509534B2 pdfaugmented reality, 3D object detection, object tracking, AR presentation, image enhancement, display technology, sensor technology, virtual reality, user interaction, environmental inputsThe disclosed invention relates to an augmented reality system capable of detecting three-dimensional objects near a user, tracking and memorizing their positions, and creating an AR presentation overlaying enhancements on the identified object, which is then displayed to the user on a device.
2020-11-03 / Blockchain-based social media history maps / US10826695B2 pdfblockchain, social media, transaction history, distributed database, trend report, chain, head node, online transaction, secure recording, immutable ledgerThe patent covers a system that integrates blockchain technology with social media by storing subscribers' online transactions as blocks within personal chains, linked to a primary or follower head node in a distributed database. The system can create new blocks from online transactions for individual subscribers and generate trend reports by analyzing data across multiple subscriber chains. It presents a novel approach to securely and immutably record social media interactions and trends within a blockchain framework.
2019-07-02 / Blockchain-based social media history maps / US10341105B2 pdfblockchain, social media, transaction history, distributed database, interaction trends, trend report, original content, cryptography, user-generated content, subscriber clusterThe patent describes a blockchain-based system for securely maintaining and linking transaction histories of social media users where each user's interactions constitute a blockchain. The system allows for the creation of trend reports based on these histories, leveraging the immutability and trustworthiness of blockchain data structures to analyze and represent user interaction trends across a network of subscribers. This decentralized approach to tracking and analyzing social media interactions can enhance the integrity and reliability of social media analytics.
2020-10-06 / Guided network management / US10797960B2 pdfvirtualized network, traffic monitoring, virtual network functions (VNFs), network snapshot, shadow network, network emulation, event mitigation, network state, network modification, service impactThis patent details a system for monitoring traffic within a virtualized network by capturing events that may impact network operations, creating snapshots to represent the network state during such events, and generating a shadow network to emulate the affected state and determine modifications to mitigate the event's negative effects.
2020-07-28 / Autonomous multi-pass data acquisition using unmanned aerial vehicles / US10726745B2 pdfautonomous data acquisition, UAV, sensor data, 3D mapping, hierarchical unit representation, high-resolution composite, multi-pass, granularity, fleet management, target areaThis patent describes an autonomous multi-pass data acquisition method for UAVs, where initial sensor data at coarse granularity is used to create a 3D hierarchical map, which then guides the UAV fleet to collect finer-granularity data for a high-resolution composite of a target area. This staged approach allows for efficient data collection by focusing on areas of interest identified in the initial survey. The method includes aggregating multi-granularity sensor data to produce an enhanced, detailed representation of the surveyed region.
2021-10-05 / Systems and methods for providing improved navigation through interactive suggestion of improved solutions along a path of waypoints / US11137257B2 pdfnavigation, dynamic routing, user interaction, waypoints, Augmented Reality, Machine Learning, real-time feedback, task integration, contextual parameters, improved navigationThe patent introduces an advanced navigation system that dynamically suggests improved routes by considering a user's location, intended destination, and diverse contextual parameters. It enhances user interaction without increasing user workload by integrating tasks and waypoints based on user preferences, proximity, and surrounding services while providing real-time feedback. The system incorporates Augmented Reality and Machine Learning to facilitate better-informed decision-making and task management during navigation.
2020-06-09 / Systems and methods for providing improved navigation through interactive suggestion of improved solutions along a path of waypoints / US10677599B2 pdfenhanced navigation, automatic routing, dynamic waypoints, user preferences, augmented reality, machine learning, real-time feedback, task management, interactive, social connectivityThe patent demonstrates a method and system for enhanced navigation by automatically suggesting improved routes with modified waypoints, dynamically adapting to user preferences and conditions such as location, proximity, and individual tasks. It integrates augmented reality and machine learning to provide real-time feedback and ancillary information, facilitating more informed decisions while en route. The system aims to reduce cognitive load and improve efficiency by managing complex task sequences and offering interactive, flexible path creation that also promotes increased social connectivity.
2021-08-31 / Determining and evaluating data representing an action to be performed by a robot / US11106284B2 pdfsensor input data, gestures, robot, action data, physical movements, processor, evaluation data, human-robot interaction, gesture interpretation, action assessmentThe invention processes sensor input data corresponding to human gestures to generate action data for controlling a robot's physical movements, with an added step of evaluating the action data against evaluation criteria to ensure the appropriateness of the robot's actions. The system involves a processor that interprets gestures into actionable commands, followed by an assessment of these commands before execution by the robot. This innovation allows for more intuitive and safe human-robot interactions by translating human gestures into robot actions while simultaneously evaluating the suitability of the actions.
2020-06-09 / Determining and evaluating data representing an action to be performed by a robot / US10678338B2 pdfsensor data, gestures, robot action, processor, physical movements, evaluation data, task performance, human-robot interaction, action determination, action evaluationThe invention pertains to a system where a processor interprets sensor data representing human gestures to determine corresponding actions for a robot to perform, which involves physical movements. The processor then evaluates the determined actions against evaluation data to ensure the appropriateness or optimization of the robot's actions. This evaluation considers factors that may influence the effectiveness or safety of the robot's performance in executing the gestures as physical tasks.
2020-03-03 / Seamless authentication device / US10581842B2 pdfauthentication system, biometric scanner, continuous authentication, authentication token, content access, seamless authentication, multi-device, user verification, biometric data, unauthorized access preventionThe patent describes an authentication system that employs biometric data for continuous user verification and generates an authentication token upon successful authentication. This token is communicated to various content providing devices, enabling seamless access to content across multiple devices without repetitive manual authentication. The system additionally addresses the issue of unauthorized content access by persistently ensuring that the user is presently consuming the content.
2021-10-19 / Time and geographically restrained blockchain services / US11151553B2 pdfblockchain, local key generation, transaction restraint, private key, preauthorization, dynamic modification, block generation, cryptographic security, immutable record, transaction conditionsThis invention presents a blockchain service device that can generate a local key with embedded transaction restraints using a private key from a user device, enhancing security and control over transactions. Upon receiving a request, the device's processor produces an input for local key generation which preauthorizes a blockchain transaction limited by the specified restraint. The system introduces the ability to dynamically modify keys with transaction conditions, which are then used to create blocks that record these preauthorized, condition-limited transactions on the blockchain.
2020-09-22 / Network controlled physical access restriction based upon physiological state / US10783727B2 pdfaccess restriction, physiological state, automated barriers, processing system, security plan, smart buildings, managed environment, detection, real-time, controlThe patent describes a system and method for controlling access to a location based on the physiological state of an individual using automatic barriers managed by a processing system, which enforces a security plan. The system detects the individual's presence and physiological state within a managed environment such as smart buildings or homes. Automated barriers are then controlled in real time according to the security plan that associates access restrictions with the detected physiological state.
2020-01-14 / Network controlled physical access restriction based upon physiological state / US10535208B2 pdfaccess restriction, physiological state, smart environment, security plan, automated barrier, smart buildings, detecting individuals, processing system, IoT, managed accessThe patent relates to selectively restricting access to an individual based on their physiological state within smart environments by employing a system that enforces a security plan. The system detects a person's physiological state within a smart environment, such as a smart building, and controls automated barriers, like doors, to grant or deny access according to the predefined security parameters. This is achieved by integrating processing systems capable of receiving security plans, detecting individuals, and assessing their physiological conditions to manage access.
2019-04-09 / Network controlled physical access restriction based upon physiological state / US10255733B2 pdfaccess restriction, physiological state, managed environment, security plan, processing system, smart buildings, automated barrier, Internet of Things (IoT), real-time, dynamic access controlThe patent describes a security system for smart environments where access restriction for an individual to certain locations is controlled based on the individual's physiological state, as defined in a security plan. The system includes a processor that detects the individual's presence and physiological condition, and operates automated barriers according to the security plan guidelines. This approach enables dynamic access control in response to real-time physiological data, enhancing security measures in managed environments.
2021-09-07 / Semantic video encoding / US11115666B2 pdfvideo encoding, theme identification, scene boundary detection, encoding strategy, thematic encoding, video program, streaming video, computing device, viewer experience, content alignmentThe technical novelty involves a method and system for encoding scenes in a video program based on identified themes, ensuring that the selected encoding strategy is thematically aligned with the content of the scene. This approach includes detecting scene boundaries, identifying a scene's theme, and choosing a theme-specific encoding strategy to apply during encoding. The encoding method enhances thematic coherence, potentially improving viewer experience and video program quality.
2019-06-04 / Audio adjustment and profile system / US10313821B2 pdfaudio system, dynamic calibration, acoustic modeling, speaker tracking, listener tracking, personalized audio, adaptive sound, user profile, mood preferences, real-timeThe patent describes an audio system capable of tracking individual listeners and speaker locations to dynamically calibrate speaker audio output for a consistent listening experience, regardless of changes in the environment or speaker/listener movement. It includes creating an acoustic model of the space to optimize speaker settings and also incorporates user profiles to tailor audio preferences, such as mood-induced frequency and tone adjustments. This adaptive audio system provides personalized and situation-aware sound calibration in real-time.
2018-05-22 / Audio adjustment and profile system / US9980076B1 pdfaudio adjustment, speaker calibration, acoustic modeling, listener tracking, audio personalization, dynamic sound environment, profile-based customization, frequency emphasis, automatic tuning, mood preferencesThe patent describes an audio system that dynamically calibrates speakers based on the tracked position of individuals and speakers within an environment, integrates an acoustic model of the room for tuning, and customizes audio playback using listener profiles, including mood-based frequency and tone preferences.
2021-12-07 / Systems and methods to augment the capacities and capabilities of cellular networks through an unmanned aerial vehicle network overlay / US11197168B2 pdfsoftware defined network, unmanned aerial vehicle, UAV, mesh network, connectivity, communication, SN device, bridging radio, LTE, UMTSThe patent describes a system leveraging a software defined network (SDN) through unmanned aerial vehicles (UAVs) equipped with SN devices and bridging radios, aimed at providing rapid deployment and expanded connectivity in varied circumstances, including emergencies. This system interconnects with existing communications infrastructures, like UMTS and LTE, to relieve saturated nodes and extend network capabilities. The UAVs form a mesh network allowing for dynamic routing decisions by a central controller, thus enhancing the coverage and robustness of the mobile network.
2018-04-17 / Systems and methods to augment the capacities and capabilities of cellular networks through an unmanned aerial vehicle network overlay / US9949138B2 pdfsoftware defined networking, UAV, wireless network, emergency communication, SDN device, bridging radio, routing decisions, RF coverage, bandwidth saturation, edge nodesThe patent introduces a system to augment pre-existing wireless networks using a UAV-based software defined network (SDN) to ensure rapid deployment and connectivity. The system bridges different communication standards like UMTS, LTE, and satellite through a UAV with an integrated SDN device and bridging radio component, managed by a central controller for optimized routing decisions. This solution aims to alleviate real-world physical constraints on networking like RF coverage issues and bandwidth saturation at edge nodes, especially in emergency scenarios or high-demand environments.
2019-11-19 / Aerial profiling of communication networks / US10484849B2 pdfsemi-autonomous vehicles, unmanned aerial vehicles, network profiling, communication networks, line-of-sight verification, radio-based problem determination, emergency service hosting, network diagnostics, dynamic evaluation, network maintenanceThe patent describes a method for using semi-autonomous vehicles, such as unmanned aerial vehicles (UAVs), to profile communication network coverage and performance issues. These vehicles perform tasks including probing and profiling airspaces, verifying line-of-sight conditions, and determining radio-based problems to support temporary hosting of services in emergency or outage situations. It improves upon traditional manual or static evaluation methods by enabling dynamic, location-specific network diagnostics and maintenance.
2018-04-10 / Aerial profiling of communication networks / US9942721B2 pdfsemi-autonomous vehicles, aerial profiling, communication networks, radio signal performance, line-of-sight verification, problem determination, temporary hosting, emergency response, network troubleshooting, dynamic analysisThe patent outlines a system for aerial profiling of communication networks using semi-autonomous vehicles, such as drones, to evaluate and enhance radio signal performance across different locations and hardware. By enabling line-of-sight verification, problem determination, and temporary hosting of services in case of emergencies, this system reduces reliance on manual inspections and fixed location reporting for network issues. The innovative approach offers dynamic, real-time analysis and troubleshooting of network coverage and performance problems.
2022-10-04 / Method and system for aggregating video content / US11462051B2 pdfvideo content aggregation, training model, image filtering, smart home, sensor data, user interface, video content adjustment, machine learning, event identification, model improvementThe disclosed system and method involve aggregating video content from premises sensors and cameras, selectively filtering this content using a training model to form an adjusted video set that includes only relevant events, and enhancing the training model based on user identifications of events within the curated content.
2020-09-29 / Method and system for aggregating video content / US10789452B2 pdfvideo content aggregation, smart home, machine learning, image categorization, content filtering, user feedback, training model adjustment, event categorization, video content analysis, model distributionThe invention relates to systems and methods for compiling video content from smart home devices, filtering it based on a machine learning model, which categorizes events and retains relevant footage while discarding non-relevant images. The content is then presented to users for identification feedback, which is used to further refine the model's accuracy in selecting pertinent categories of events. Additionally, the improved model can be distributed to network devices for enhanced video content aggregation and analysis.
2019-06-18 / Method and system for aggregating video content / US10325143B2 pdfvideo content aggregation, smart home, training model, image filtering, machine learning, content presentation, user interaction, model refinement, network devices, content adjustmentThe patent describes systems and methods for aggregating video content from smart home devices, and selectively adjusting this content based on a training model which filters images into relevant and non-relevant subsets. The relevant subset, determined by the machine learning model in relation to various events, is presented to the user for identification, which in turn refines the training model. This refined model is then provided to network devices to enhance future content filtering and aggregation.
2018-03-13 / Method and system for aggregating video content / US9916493B2 pdfvideo content aggregation, smart home, training model, image filtering, content presentation, user feedback, machine learning, event categorization, video footage review, adaptive systemThe invention provides a method for aggregating video content from smart home devices, filtering it based on a training model to present a refined subset of images relevant to specific events, and improving the model through feedback on the identified images. This system refines the aggregation of video footage by selectively including content categorized by event types, thereby enhancing user review efficiency. The method also includes an iterative learning process where the training model adapts to user-provided identifications, improving subsequent content filtering and presentation.
2020-03-17 / Content quality assessment and prediction via flows / US10594569B2 pdfquality of service, communication network, content flow model, flow statistics, network nodes, anomaly detection, content delivery path, Deep Packet Inspection, service quality assurance, adaptive network managementThe patent details a method for improving quality of service in communication networks by constructing a model of content flow from aggregated flow statistics, predicting operations of network nodes based on the model, and adjusting content flow on delivery paths in response to detected network anomalies. It focuses on detecting and responding to service quality issues without Deep Packet Inspection by comparing detected quality of service with predictions from the model. The invention also involves anomaly detection and content delivery path adjustment to maintain service quality.
2018-11-27 / Content quality assessment and prediction via flows / US10142185B2 pdfcontent flow data, communication network, flow statistics, predictive model, quality of service, network node operation, anomaly detection, dynamic adjustment, content delivery path, service assuranceThe patent details a method of analyzing content flow data in communication networks to derive flow statistics for constructing a predictive model, which is then used for predicting network node operations and identifying anomalies by comparing detected and predicted quality of service, thereby enabling dynamic adjustment of content flow based on anomalous activity detection.
2022-11-22 / Automated content selection for groups / US11509958B2 pdfmedia content selection, venue, aggregated content profile, viewer preferences, location data, profile data, conflict resolution, weighting preferences, personalization, dynamic adjustmentThe invention pertains to a system and method for personalized media content selection for groups in a venue, utilizing location and profile data to create an aggregated content profile. On detecting approaching viewers with imaging data, it resolves content preference conflicts by weighting individual preferences. This tailored content delivery approach enhances group media experiences by dynamically adjusting to the viewers' profiles.
2021-06-15 / Automated content selection for groups / US11039210B2 pdfmedia content selection, viewer preferences, profile data, aggregated content profile, conflict resolution, location data, personalization, venue display, weighted preferences, group viewingThe patent discloses a system and method for selecting media content at a venue by aggregating the content viewing preferences of a group of persons based on their location and profile data. It resolves conflicts in viewing preferences with a weighted system when a new group approaches the display device. The technology aims to enhance the viewing experience through personalization and anticipation of group preferences.
2020-02-11 / Automated content selection for groups / US10560745B2 pdfmedia content selection, venue, viewer location data, profile data, aggregated content profile, dynamic content generation, weighted preference resolution, personalized media delivery, display device, collective viewing preferencesThe patent describes a system and method for media content selection at a venue, utilizing location and profile data of viewers to create an aggregated content profile. It dynamically selects content based on the aggregated preferences and resolves conflicts through weighted preferences as viewers approach a display device. The technology anticipates personalized media delivery based on the collective characteristics of viewers in a shared physical space.
2018-10-02 / Automated content selection for groups / US10091550B2 pdfmedia content selection, group preferences, venue, location data, profile data, aggregated content profile, imaging data, real-time adjustment, conflict resolution, preference weightingThe patent describes a system and method for automatically selecting media content tailored to the preferences of a group of people at a specific venue using location and profile data to aggregate viewing preferences and resolve conflicts. The approach uses imaging data to sense the presence and approach of viewers, adjusting content selection in real-time. Weighting mechanisms are applied to balance the group's conflicting content preferences to optimize the viewing experience.
2022-08-02 / Method and apparatus for providing a recommended action for a venue via a network / US11405486B2 pdfrecommended action, perception model, sensor data, venue, user feedback, real-time adjustments, environmental control, data analysis, user experience, threshold-basedThe invention relates to a method and apparatus that collect data from sensors in a venue and user devices to inform a perception model, which then generates and updates recommended actions when certain thresholds are met. This allows for real-time adjustments in venue services based on environmental factors and user feedback. The system provides dynamic control of venue settings, improving customized user experiences and venue operations.
2020-07-28 / Method and apparatus for providing a recommended action for a venue via a network / US10728363B2 pdfrecommended action, sensor data, venue management, perception model, user feedback, environmental control, adaptive system, threshold, real-time updates, user devicesThe patent introduces a method and system that utilizes both environmental sensors at a venue and sensors in user devices to gather data, which is then applied to a venue perception model. A recommended action is generated once the combined data meets a predetermined threshold, and updates to the perception model are implemented based on this action. The result is an adaptive and responsive recommendation system to optimize venue management and user experience by integrating direct user feedback.
2019-03-05 / Method and apparatus for providing a recommended action for a venue via a network / US10225369B2 pdfrecommended action, sensor data, venue management, perception model, dynamic adjustment, user input, environmental control, data integration, threshold, responsive systemThe patent introduces a system that integrates sensor data from a venue and user devices to inform a recommendation engine that generates actions once certain thresholds are met, with continuous updating of the venue's perception model. This system takes user input into account, providing dynamic and responsive venue management, such as environmental control adjustments. The process includes the collection of data through venue and user device sensors, application of a perception model, and provision of updates to this model influenced by the generated recommendations.
2019-09-03 / Applying machine learning to heterogeneous data of existing services to generate a new service / US10404549B2 pdfmachine learning, policy generation, data modeling, new service creation, multi-service analysis, policy models, service functionality implementation, competitive advantage, market need analysis, risk mitigationThe patent describes a method and system for creating new services by utilizing machine learning to derive policies from multiple pre-existing services and combining these policies to implement functions of the new service. This process includes generating distinct policies from different services using respective policy models and then applying these policies to establish the functionalities of a new service that potentially fulfills market needs. The approach aims to mitigate the risk and costs traditionally associated with launching untested services while providing a competitive advantage through timely introduction of the new service.
2020-09-08 / Billboard-based advertising system / US10769667B2 pdfdynamic advertising, vehicle detection, demographic profiling, targeted messaging, broadcast system, on-board device, roadside billboards, personalized advertisement, message generation, communication systemsA system identifies a vehicle and its occupant's demographic profile, then selects and broadcasts targeted advertising messages directly to the vehicle's onboard device. The system overcomes the limitations of static roadside billboards by providing dynamic, demographic-specific advertisements. Operations involve vehicle detection, demographic profiling, advertisement tailoring, message generation, and transmitting the tailored message to the on-board device.
2022-02-15 / Vehicle-based mobile node fleet for network service deployment / US11252638B2 pdfservice assignment, vehicle-based mobile node, network function virtualization, software-defined networking, mobile endpoint device, route determination, communication network, SDN, NFVI, telecommunication resourcesThe patent discloses a method for dynamically assigning telecommunication services to a vehicle-based mobile node by determining the coincidence of routes between the mobile endpoint device requesting the service and the vehicle-based node. The process utilizes a software-defined network (SDN) and network function virtualization infrastructure (NFVI) to efficiently provision network services. This capability allows for adaptable and efficient use of telecommunication network resources based on the mobility patterns of endpoint devices and vehicle nodes.
2019-09-03 / Vehicle-based mobile node fleet for network service deployment / US10405260B2 pdfnetwork function virtualization, NFVI, vehicle-based mobile node, dynamic service allocation, route determination, mobile endpoint, SDN architecture, reconfigurable nodes, communication network, service assignmentThe patent describes a system and methods for dynamically allocating network services to reconfigurable vehicle-based mobile nodes by determining coinciding routes between a mobile endpoint device and the vehicle-based node. The system utilizes a processor within a communication network to evaluate service requests and routes, effectively leveraging network function virtualization infrastructure (NFVI). This enables adaptable assignment of network functions and services, optimizing the use of shared network resources.
2022-05-10 / Optimization of multiple services via machine learning / US11328224B2 pdfmachine learning, policy generation, service optimization, data modeling, adaptive policies, service provider, multi-service application, policy transfer, dynamic conditions, policy implementationThis patent details a method for transferring and applying policies between services using machine learning models to adapt to changing conditions, enabling the generation of a new policy for a second service based on the policy of a first service, provided that both services are under the same service provider. The system assesses the applicability of a generated policy for one service to another and applies it to modify or create a new policy via another machine learning model. This process improves service optimization by adapting policies to dynamic service environments and customer needs.
2020-07-21 / Optimization of multiple services via machine learning / US10719777B2 pdfmachine learning, policy model, data modeling, service optimization, policy generation, policy implementation, dynamic adaptation, service provider, cross-service application, autonomous systemThe innovation provides a method and system for generating and applying policies across different services using machine learning models. It involves creating a policy for one service based on its data and machine learning, assessing its applicability to another service, and, if relevant, applying and adapting the policy to improve the second service. This process is managed by a single service provider, enabling dynamic policy generation and implementation under changing conditions.
2020-07-14 / Generating a pre-caching schedule based on forecasted content requests / US10715959B2 pdfmedia content, wireless communications, base station, pre-caching, multicast transmission, user devices, content forecast, network congestion, media consumption data, stream interruptionsThe patent describes a base station managed method for reducing data stream interruptions by pre-caching media content. It forecasts media content requests from user devices and sends multicast transmissions to multiple devices, with instructions for first devices to store content for later use based on forecasted demand. This pre-caching approach minimizes busy period network congestion and enhances user experience by avoiding interruptions during peak traffic times.
2018-10-30 / Generating a pre-caching schedule based on forecasted content requests / US10117058B2 pdfmedia data delivery, mobile devices, pre-caching, forecasted content requests, expected locations, multicast transmission, user experience, streaming optimization, data stream interruption, content pre-caching scheduleThis invention is a method for optimizing media data delivery to mobile devices by pre-caching media content based on forecasted content requests and expected device locations. It involves generating a pre-caching schedule and issuing a multicast transmission to multiple devices, minimizing data stream interruptions and improving user experience. The method includes sending instructions to devices for storing media data, and responding to subsequent requests for the same media via multicast transmission.
2021-03-02 / Agent for learning and optimization execution / US10936966B2 pdfautonomous agent, dynamic system, model updating, minimum acceptable criteria, performance evaluation, reward modeling, model improvement actions, user-defined standards, system maintenance, adaptive learningThe patent describes a self-improving autonomous agent designed for dynamic systems that monitors and updates an underlying model based on minimum criteria set by users. The agent continually compares model outputs against these criteria, and when discrepancies arise, it generates and evaluates potential improvements by estimating their rewards. The agent then executes the action with the highest calculated reward to enhance model performance.
2021-05-25 / Video motion augmentation / US11019393B2 pdfvideo motion augmentation, micro-movements, content delivery networks, device characteristics, sensor data, user feedback, display resolution, viewing experience, content adaptation, video enhancementThe patent describes a system for enhancing video motion specifically by improving visibility of small or micro-movements on devices with various display capabilities. This augmentation is influenced by factors such as the device characteristics, content theme, sensor data, and user feedback to tailor the video playback experience. The system also aims to address issues with viewing experience on small devices or in environments with excessive external movement, such as watching content on a train.
2019-10-15 / Video motion augmentation / US10448094B2 pdfvideo motion augmentation, micro-movements enhancement, content delivery networks, device characteristics, user feedback, sensor data, content theme, viewing experience, adaptability, dynamic enhancementThe disclosed innovation provides a system to dynamically enhance micro-movements in video content for varying device displays and user environments. It analyzes video content and metadata to adapt these small movements based on the device characteristics, content theme, sensor data, and user feedback. The enhancement aims to improve viewing experiences on devices with different capabilities and under varying motion conditions, such as while being on a moving train.
2018-01-30 / Video motion augmentation / US9883235B2 pdfvideo motion augmentation, micro-movements enhancement, content delivery network, device characteristics, adaptive video enhancement, sensor data utilization, user feedback, display resolution, environmental motion compensation, quality of experienceThe patent describes a system for dynamically augmenting and emphasizing small or micro-movements in video content, tailored to various inputs including device characteristics, content theme, sensor data, and user feedback. Enhancements are made to improve visibility and quality of experience on devices with varying display capabilities, including compensating for environmental motion that may affect viewing. The system utilizes content delivery networks to provide efficient and improved video content quality to users across different devices and conditions.
2020-09-22 / Telemedicine application of video analysis and motion augmentation / US10779733B2 pdftelemedicine, video analysis, motion augmentation, anomaly detection, non-invasive, medical diagnosis, patient interaction, alerts, iterative analysis, portable systemThe patent describes a non-invasive telemedicine system that employs video analysis to detect both macro and micro anomalies in a patient's movements to aid in medical diagnosis. Upon identification of an anomaly, the system issues an alert and formulates a request for interaction to gather further diagnostic information. Through iterative analysis, the system refines its diagnosis without the need for traditional, invasive medical equipment or expensive, non-portable devices.
2021-12-07 / Advertisement generation based on a user image / US11197061B2 pdfcustomized multimedia content, image analysis, targeted advertising, content personalization, viewer context, computing device, user equipment, background component selection, foreground characteristic, image data processingThe invention is a method for creating customized multimedia content, wherein a computing device receives an image of a viewer and then selects a background component for the content based on an identified item in the image. This personalized content is then provided to user equipment for display to the viewer, integrating the selected background with a foreground component that possesses a specific characteristic. The system aims to generate individualized advertisements, enhancing targeting by tailoring content to the viewer's context as captured in the image data.
2020-10-13 / Advertisement generation based on a user image / US10805678B2 pdfmultimedia content, image data, personalization, targeted advertising, viewer profiling, dynamic content integration, computing device, user equipment, content display, advertisement customizationThe patent details a method wherein a computing device selects multimedia content components based on image data of a viewer, with the aim of personalizing content such as advertisements. By identifying characteristics within the image data, a more targeted background component is dynamically integrated into the multimedia content. This content is then provided to user equipment for display to the viewer, enhancing the relevance of the advertisement to the individual viewer.
2018-07-24 / Advertisement generation based on a user image / US10034050B2 pdfuser response, image analysis, advertisement customization, foreground component, background component, personalized advertising, dynamic content generation, targeted marketing, user interaction, ad customizationThe patent describes a method for customizing advertisements by analyzing an image of a user's reaction to an initial ad, selecting a new foreground component corresponding to the user's response, and combining it with the existing background component to create an updated, personalized advertisement.
2020-04-21 / Automated obscurity for digital imaging / US10628922B2 pdfprivacy, obfuscation, digital media, smartphone, flag, image capturing, face obscuring, mobile device, user consent, image processingThe patent outlines a privacy-protection technology whereby individuals can transmit a flag from their mobile device indicating a desire to have their image obscured in any digital media captured by other devices. Upon capturing digital media, the capturing device receives this flag and processes the media by obfuscating the flagged individual's face or entire image. The innovation ensures privacy by actively influencing image or video capture processes in devices within the vicinity of the user's smartphone.
2017-06-13 / Automated obscurity for pervasive imaging / US9679194B2 pdfobfuscation, privacy protection, captured media, facial features, motion information, image detection, endpoint device, media recording, feature set, subject-initiated communicationThe patent describes a method for preserving privacy by obfuscating images of subjects in captured media, using a subject-initiated communication containing a feature set to detect and then obfuscate the subject's image. The feature set comprises facial features and motion information to accurately match and track the subject in the media. Upon detection, the subject's image is actively obfuscated to protect privacy in environments where pervasive recording devices are used.
2020-08-11 / Object based image processing / US10740885B2 pdfimage processing, object quality value, blurriness value, color value, object category metrics, image quality assessment, comparative results, quality improvement, image enhancement, multiple objectsThis method enables tailored image processing by determining individual object quality values—such as blurriness and color—for multiple objects within an image and comparing them to predefined object category metrics. Different image quality assessment and processing operations are then applied based on the comparative results for each object, allowing for object-specific quality improvement. The innovation allows for more granular and accurate image enhancement in scenarios where images contain multiple objects with varying quality levels.
2018-11-27 / Object based image processing / US10140696B2 pdfimage processing, object quality values, object category metrics, image enhancement, quality assessment, image correction, image data, individual objects, comparison, adaptive processingThe patent describes a method for image processing that determines object quality values for individual objects within an image and compares these to associated object category metrics. This enables the adjustment of image processing techniques based on the quality assessment of each object, rather than applying a global or uniform approach. The innovation allows for sophisticated, object-specific image enhancement or correction by accounting for the varying quality of different objects within the same image.
2017-02-14 / Object based image processing / US9569832B2 pdfimage processing, object quality, image alteration detection, object category metrics, notification generation, quality assessment, fine-grained detection, image tampering, multiple objects, clarity discrepancyThe patent describes an image processing method that detects alterations by comparing quality values of specific objects within an image to predefined object category metrics, triggering a notification if a discrepancy indicative of alteration is found. It improves upon previous methods by providing quality assessments on an object-by-object basis, rather than globally for the entire image or merely locally within portions, allowing for fine-grained detection of image tampering. This technique is especially beneficial in scenarios where multiple objects of varying clarity are present within a single image.
2016-07-19 / Object based image processing / US9396409B2 pdfobject detection, image processing, quality assessment, image data, object quality values, object category, comparison, category metrics, image enhancement, action initiationThe method detailed in the patent improves object-specific image quality assessment by detecting objects within an image, determining quality values for a selected image data portion corresponding to each object, and comparing those values to predefined metrics of different object categories. This comparison triggers appropriate actions based on whether the object's quality meets the category-specific standards. The system caters to images with multiple objects of varying quality, providing a more nuanced quality detection mechanism than conventional whole-image or segment-based approaches.
2021-02-16 / Image quality estimation using a reference image portion / US10922580B2 pdfimage quality estimation, feature comparison, pixel subsets, common feature identification, scene analysis, variation analysis, image comparison, quality evaluation method, photograph analysis, image processingThis invention introduces a method for determining image quality by identifying and comparing features between a pair of images of the same scene to estimate quality based on variation in corresponding pixel subsets. The method uses a comparison of variation within pixel subsets identified from common features in different images rather than relying on full, reduced, or no image references. This approach enables quality estimation by assessing variances between specific aligned portions of images rather than whole images or unrelated references.
2018-07-17 / Image quality estimation using a reference image portion / US10026010B2 pdfimage quality estimation, pixel subset comparison, multiple images, common scene, reference methods, image processing, best quality identification, photographs, partial reference, computing deviceThe technical novelty involves a method for estimating image quality by comparing subsets of pixels between two images of the same scene, differing from conventional full or reduced reference methods which compare entire images or an image to a reference portion. This pixel subset comparison allows for an image quality assessment without requiring the full reference image or a defined portion, providing more flexibility and potential efficiency in processing. The concept is designed to identify the image with the best quality in a set through a targeted analysis of corresponding pixel subsets between multiple images.
2020-12-08 / Method and apparatus for organizing media content / US10860886B2 pdfimage clustering, similarity matrices, common features, processor-based system, quality-based filtering, target cluster image, media content management, automatic organization, image categorization, cross-device experienceThe patent describes a method for automatically organizing a collection of images into cluster groups determined by a processor-based system, which identifies common features of the images using similarity matrices. Cluster groups are defined containing images that all share these common features, and a quality-based filtering is applied to select a representative target cluster image for each group. This technology enhances media content consumption by efficiently categorizing and summarizing a large number of images for improved user experience across different devices.
2018-06-19 / Method and apparatus for organizing media content / US10002310B2 pdfimage clustering, similarity matrices, common features, quality-based filtering, cluster groups, processor, media content, representative image, automatic organization, image collectionsThe patent details a method for organizing a large number of images into clusters using similarity matrices to determine common features across images, defining cluster groups with these common features, and applying quality-based filtering to select a representative image from each cluster. The system leverages a processor to automate the organization and representation of images which can enhance users' experience with media content on various devices. It provides a structured approach to manage and navigate through extensive image collections efficiently.
2021-03-09 / Method and apparatus for augmenting media content / US10945035B2 pdfnarrative descriptions, media content, synchronization, timestamps, audio reproduction, visual impairment, content analysis, accessibility, cross-device, enhanced media experienceThe patent relates to a method for automatically generating narrative descriptions of media content, such as visual features and events, syncing these narrations with timestamps in the media content, and presenting an audio version of these descriptions in synchronization with the media content's playback. This system provides a tailored media experience across multiple devices by analyzing media content to create descriptive audio tracks for visual content. The technology enhances accessibility for consumers, including those with visual impairments, by aligning descriptive audio narrations with corresponding visual events in the media content.
2019-09-17 / Method and apparatus for augmenting media content / US10419818B2 pdfnarrative generation, media content, audio description, visual features, synchronization, time stamps, accessibility, content consumption, communication systems, web portalThe patent relates to a method for generating and synchronizing narrative descriptions with media content, where the narrative is based on visual features and events within the content and includes time stamps for accurate alignment with the media's video playback. This enables enhanced media content consumption by providing audio descriptions synchronized with visual content, improving accessibility for users across various devices. The generated narrative descriptions can be configured and managed through specialized web portals and communication systems.
2017-09-19 / Method and apparatus for augmenting media content / US9769524B2 pdfnarrative descriptions, media content, visual features, timestamps, audio reproduction, synchronization, enhanced consumption, system architecture, communication systems, web portal configurationThe disclosed system generates narrative descriptions for media content by analyzing visual features and events, timing them with timestamps. These narrative descriptions are presented as audio reproductions synchronized with the video content, enhancing media consumption across various devices. Multiple figures depict the system architecture, narrative generation process, communication systems involved, and configurations through web portals and communication devices.
2016-09-20 / Method and apparatus for augmenting media content / US9451335B2 pdfnarrative descriptions, media content, visual features, audio reproduction, time stamps, synchronization, video playback, device compatibility, media consumption, content analysisThis technology generates synchronized narrative descriptions for media content by analyzing visual features and events, and pairs these descriptions with corresponding time stamps. The narratives are then presented as an audio reproduction synchronized with the media content video playback. This system especially caters to a variety of devices and offers enhanced media consumption experiences.
2021-04-13 / Delivery of media content to a user device at a particular quality based on a personal quality profile / US10979755B2 pdfpersonal quality profile, media content customization, server-side modification, visual attributes, media quality preferences, content adaptation, user device, automatic adaptation, content delivery, personalizationThe technical novelty lies in a system and method for delivering customized media content to a user's device by modifying the visual attributes of the content at the server-level based on a user's personal quality profile when the preferred version is not available. This personal quality profile reflects a user's media quality preferences, allowing for automatic adaptation of media content to suit individual viewing requirements. The server modifies a retrieved version of the content to generate the personalized version, which is then sent to the user's device.
2019-05-21 / Delivery of media content to a user device at a particular quality based on a personal quality profile / US10298987B2 pdfmedia streaming, personal quality profile, user device, server computing device, media quality preferences, delivery bit rate, content adjustment, user preferences, streaming method, media content versionThe technical novelty involves a server computing device streaming media content to a user device based on a retrieved personal quality profile that indicates the user's media quality preferences and adjusts the media content version sent to accommodate for both personal preferences and delivery bit rate.
2020-10-13 / Method and apparatus for providing media content / US10805657B2 pdfmedia content, high resolution, bandwidth optimization, resolution threshold, service provider network, unused bandwidth, communication systems, content streaming, media server configuration, ultra high definitionThe patent discloses a method for efficiently utilizing bandwidth by receiving media content at an initial high resolution, determining if this resolution exceeds a viewing threshold, then notifying the service provider to adjust the resolution to a lower level, thereby freeing bandwidth to receive additional media content over the same channel. This approach allows for simultaneous transmission of multiple streams of media content over a high-definition channel by optimizing resolution and bandwidth utilization. It also includes systems and configurations for implementing the method, including communication devices and a web portal for media server configuration.
2018-01-30 / Method and apparatus for providing media content / US9883222B2 pdfadaptive media delivery, high-resolution channel, resolution threshold, unused bandwidth, content reallocation, dynamic resolution adjustment, bandwidth optimization, media content streaming, service provider network, multiple media contentThe technical novelty involves a method for adaptive media content delivery over high-resolution channels by determining if the initial content resolution exceeds a threshold, subsequently adjusting that resolution upon notification to free bandwidth, and utilizing the freed bandwidth to receive additional media content. This optimizes network resource usage by dynamically and efficiently reallocating unused bandwidth for supplementary content. The method enables simultaneous enjoyment of multiple media streams on a single channel without sacrificing the quality of primary content.
2016-12-06 / Method and apparatus for providing media content / US9516358B2 pdfmedia content delivery, bandwidth optimization, resolution threshold, content rescaling, UHD channel, service provider network, high resolution media, unused bandwidth, multiple devices, consumer media experienceThe patent describes a method for optimizing media content delivery by adjusting the resolution of transmitted media content to free up bandwidth when the initial resolution exceeds a viewing threshold. Upon notification of the excessive resolution, the content provider rescales the content to a lower resolution, which enables the simultaneous transmission of additional media content over the newly available bandwidth. The system enhances consumer media experiences by allowing a variety of content to be delivered over multiple devices via a single ultra high definition (UHD) channel.
2020-10-20 / Method and apparatus for generating information associated with a lapsed presentation of media content / US10812875B2 pdfmedia program, content summary, missed portion, segment identification, optimal timing, presentation initiation, non-disruptive, user-specific, media services, viewing experienceThe invention enables provision of summaries for missed content portions in media programs without disrupting ongoing viewing, by identifying the relevant content segment, scheduling an optimal time to present the summary, and initiating the presentation to the user who missed the original content.
2018-07-24 / Method and apparatus for generating information associated with a lapsed presentation of media content / US10034065B2 pdfmedia program, summary presentation, content identification, optimal timing, automatic triggering, shared television system, non-disruptive, catch-up feature, missed content, user experienceThe patent introduces a system for providing summaries for missed content in a media program without disrupting ongoing viewing, by identifying unwatched segments and determining optimal times for summary presentation based on content subject matter. The summary presentation is triggered automatically to help a user catch up on missed content. This system avoids the need for replaying the entire program and thereby does not interrupt other viewers.
2017-10-31 / Method and apparatus for generating information associated with a lapsed presentation of media content / US9807474B2 pdfmedia program, summary generation, missed content, subject matter identification, presentation timing, non-interruptive playback, content recap, optimal presentation timing, viewer pattern, media consumptionThe patent relates to generating tailored summaries of missed content in live media programs without interrupting other viewers, by identifying key subject matter in the missed segment and scheduling a summary presentation at an opportune time. It prevents playback disruptions for viewers who are up-to-date with the media program by allowing individuals who missed parts of the program to catch up seamlessly. The system determines the optimal moment for recap presentation based on the content's context and viewer's media consumption pattern.
2021-03-09 / Manipulation of media content to overcome user impairments / US10943502B2 pdfmedia content modification, impaired perception, real-time, personalized media consumption, simultaneous presentation, content adaptation, viewer preferences, multimedia experience, electronic devices, media presentation deviceThe invention identifies impairments in the perception of original media content and determines a modification to create a modified media content version that improves the user's experience. These modifications are applied in real time to allow simultaneous presentation of the original and modified content for viewers with different viewing preferences or needs. The system enables personalized media consumption on various devices without altering the original media content for all viewers.
2019-11-26 / Manipulation of media content to overcome user impairments / US10490099B2 pdfmedia content modification, impaired perception, simultaneous presentation, consumer experience, personalized viewing, multimedia compensation, media presentation device, original content, enhanced media, device settingsA system is provided that identifies conditions impairing the perception of media content and generates a companion modified media item to be presented simultaneously with the original content, enhancing the consumer experience regardless of individual differences or device settings. The modified media content compensates for the perceived impairment, allowing viewers with different preferences or experiencing different conditions to enjoy the media content without altering the original presentation. This simultaneous presentation of original and modified media content items enables personalized viewing experiences shared across diverse audience members.
2019-11-05 / Method and apparatus for providing secure image encryption and decryption / US10467427B2 pdfimage decryption, partial decryption, encrypted image display, decryption key, selective display, user authorization, confidential documents, adjacent portion decryption, encryption keys, privacy protectionThe patent describes methods and systems for decrypting only a portion of an encrypted image using a decryption key and using partial information from the decrypted portion to display an adjacent encrypted portion. It also details a method to decrypt and display a first image portion using a first key, followed by a second encrypted portion using a second key. These methods ensure the confidentiality of sensitive documents by providing selective display of encrypted documents based on user authorization.
2016-07-19 / Method and apparatus for providing secure image encryption and decryption / US9396310B2 pdfpartial decryption, encrypted image, decryption key, adjacent image information, confidentiality, dual-key encryption, document security, compartmentalized access, privacy, image displayThe technical novelty includes methods and systems that enable partial decryption and display of encrypted images, utilizing decryption keys to reveal only the section that the user is entitled to see, while adjacent sections can be inferred using partial information. A dual-key encryption method is also disclosed, where a first portion of an image is decrypted with a first key for display, enhancing confidentiality in document handling. The invention ensures that sensitive information is compartmentalized, reducing the risk of unauthorized access to entire documents or images.
2021-03-30 / Mobile service platform / US10963822B2 pdfcrowdsourcing, service platform, task allocation, worker device, device attributes, task completion, data validation, customer interaction, rewards system, real-time updatingThe patent describes a crowdsourcing-based service platform which allocates tasks to registered worker devices based on their current attributes, validates the results, and relays data back to the customer, thereby streamlining the task completion process and potentially incentivizing the participating devices.
2018-01-23 / Mobile service platform / US9875448B2 pdfservice platform, device-based task completion, attribute-based worker device selection, data validation, customer presentation, worker device incentive, crowdsourcing, task allocation, device-specific capabilities, performance incentivesThe disclosed invention details a service platform that selects a worker device from a pool of registered devices to complete tasks based on the device's current attributes. Upon task completion, the collected data is verified and then provided to the requesting customer, with the device potentially receiving a reward or incentive for successful task fulfilment. This method optimizes task allocation and completion in crowdsourcing by leveraging device-specific capabilities and performance incentives.
2020-05-19 / Opportunistic crowd-based service platform / US10659527B2 pdfcrowd-based service, mobile sensor devices, data aggregation, opportunistic, location-based, user demographics, content delivery, data integration, tailored content, sensor dataThe patent introduces a method and apparatus to leverage opportunistic crowd-based service platforms, utilizing mobile sensor devices selected based on location, intrinsic properties, and user demographics. Data from these devices is compiled and integrated with information from other sensors to generate aggregated data. This processed information is then utilized to deliver tailored content to user devices.
2018-11-20 / Opportunistic crowd-based service platform / US10135920B2 pdfopportunistic crowd based service, mobile sensor device, data aggregation, content delivery, user demographics, targeted services, data collection, crowdsourcing, sensor data, service platformThis patent describes a system that selects mobile sensor devices based on specific criteria, collects data from these devices, and aggregates it with data from other sensors. The aggregated data is then used to generate content that is delivered to user devices. This method improves crowd based service platforms by leveraging opportunistic data collection and user demographics to create targeted services.
2018-01-30 / Opportunistic crowd-based service platform / US9882978B2 pdfopportunistic crowd based service, mobile sensor device, data aggregation, crowdsourcing, user demographic data, content delivery, sensor data collection, creative solutions, network of devices, cost-effectiveThe patent details a system for an opportunistic crowd based service platform that selects mobile sensor devices for data collection based on their location, properties, or user demographic data. This data is then combined with data from other devices to create aggregated content, which is subsequently delivered to users. The process utilizes crowd-sourcing principles to provide creative and cost-effective solutions by leveraging a network of user devices as data sources.
2017-02-21 / Opportunistic crowd-based service platform / US9578095B2 pdfcrowdsourcing, mobile sensor device, data aggregation, opportunistic service platform, content delivery, user demographics, sensor data, location-based services, data collection, targeted content generationThe patent describes a system that selects mobile sensor devices based on location and other attributes to collect data, which is then aggregated with data from other sensors. The aggregated data is used to generate content that is delivered to user devices, facilitating an opportunistic crowd based service platform. Different user or sensor demographics and historical data enhance the service by targeting the most relevant data sources.
2015-06-16 / Opportunistic crowd-based service platform / US9058565B2 pdfcrowdsourcing, mobile sensor device, data aggregation, opportunistic service, crowd based, data collection, sensor data, content delivery, user device, service platformThis patent introduces an opportunistic crowd-based service platform that employs mobile sensor devices selected by location and/or user qualities. The platform collects data from these devices, aggregates it with information from other sensors, and delivers generated content based on this collective data to users. It leverages crowdsourcing to provide innovative solutions by analyzing aggregated sensor data for various applications.
2020-03-03 / System and method to digitally replace objects in images or video / US10580219B2 pdfdynamic product placement, video manipulation, object replacement, advertising, 3D modeling, video content, computer system, visible feature, seamless integration, demographic targetingThe patent outlines a computer-implemented method for replacing an object in a video frame with an alternate object by detecting the absence of a visible feature, manipulating a 3D model of the new object to include the feature, and inserting the generated representation into the video. This technique allows for dynamic product placement and tailored advertising without the need for reshoots, facilitating seamless integration of updated or demographic-specific product representations into existing video content. The method leverages a system designed to automate the process of object replacement in videos, enhancing the relevance and effectiveness of product placement in media.
2019-04-02 / System and method to digitally replace objects in images or video / US10249093B2 pdfvideo frame, object replacement, 3D model manipulation, visual property matching, dynamic product placement, advertising, content adaptation, video content, seamless integration, viewing demographicsThe technical novelty lies in a method for replacing objects within a video frame, where the replacement is based on rules matching objects and the manipulations of a 3D model to match visual properties of the original object. The method ensures that the replaced object integrates seamlessly with the existing visual context by adjusting its visual properties accordingly. This process allows for dynamic product placement in videos to align with changing demographics or product designs without the need for reshoots.
2017-04-18 / System and method to digitally replace objects in images or video / US9626798B2 pdfdynamic product placement, video content, object identification, object replacement, three-dimensional modeling, visual matching, advertising, content manipulation, demographic targeting, design updatesThis method facilitates dynamic product placement in video content by identifying an object within a video frame and determining whether it should be replaced based on predefined rules. Upon deciding to replace the object, a three-dimensional model of the new object is manipulated to visually match the original object's properties. The original object in the video frame is then replaced with the modified model of the new product to tailor the advertisement to different demographics or updated product designs.
2014-10-21 / Telepresence simulation with multiple interconnected devices / US8868684B2 pdftelepresence, multimedia communication, interconnected devices, external controller, network, media information, presence simulation, selective enabling, device coordination, shared spaceThis patent describes a method for coordinating telepresence by selectively enabling and disabling multimedia communications on interconnected devices in a shared space using an external controller across a network. It includes determining the presence and multimedia capabilities of devices within a common space, and controlling media information communication type and device usage to simulate presence. The telepresence system ensures that only one device is actively communicating a specific type of media at a time, to enhance the telepresence simulation experience.
2018-07-24 / Dynamic access to external media content based on speaker content / US10031651B2 pdfinteractive conference, supplemental media content, terminology monitoring, external sources, real-time, audio/video conference, supplemental material, contextually relevant, dynamic access, enriched conferencing experienceThis patent describes a system that enhances audio/video conferences by dynamically providing supplemental media content based on words selected from the terminology used within the conference. The system monitors the terminology content from devices in an interactive conference, identifies relevant supplemental media from external sources, and makes it available to participants. The technology allows for an enriched conferencing experience by integrating contextually relevant external media in real-time.
2015-09-01 / Dynamic access to external media content based on speaker content / US9124660B2 pdfinteractive conference, terminology monitoring, supplemental media content, dynamic identification, external sources, audience enrichment, content presentation, media coordination server, network, participant engagementThe patent describes a system for enhancing interactive conferences by monitoring terminology used by participants and using this information to dynamically identify and provide supplemental media content from external sources. The supplemental content is selectively made available to audience members to enrich the conference experience. This process is facilitated by a combination of monitoring, media coordination servers, and networks.
2014-05-06 / Dynamic access to external media content based on speaker content / US8719031B2 pdfaudio conference, speaker content, supplemental media, external source, audience member, content monitoring, media presentation, dynamic access, computer system, real-timeThis patent describes a system for enhancing audio conferences by monitoring speaker content, selecting relevant words, and dynamically providing associated supplemental media content from external sources to audience members. Supplemental media is identified and displayed to participants to enrich the conference experience. The system utilizes a computer with a processor and memory for the selection and presentation process.
2019-06-04 / Speaker association with a visual representation of spoken content / US10311893B2 pdfaudio conference, speaker content, audience profile, word weighting, content personalization, visual representation, participant interests, dynamic adjustment, monitoring spoken content, selective presentationThe invention monitors spoken content during an audio conference and applies different weights to words based on audience member profiles to determine a relation between speaker content and each audience member's profile. Visual representations of the speaker content are then presented selectively to audience members based on the determined relation, enhancing the personalization and relevance of the content. This technology accommodates individual participant interests and preferences in an audio conference by dynamically adjusting the visual content.
2017-08-29 / Speaker association with a visual representation of spoken content / US9747925B2 pdfaudio conference, visual representation, speaker content, audience profile, word classification, content weighting, personalization, monitoring, relation determination, customized visualizationThis patent relates to a method for customizing visual representations of speaker content during audio conferences by classifying words based on audience member profiles. Weight is assigned to words in accordance with parameters from each audience member's profile. Consequently, this allows for the presentation of personalized visual content to different audience members based on the relevance of the speaker's words to their individual profiles.
2017-04-04 / Speaker association with a visual representation of spoken content / US9613636B2 pdfaudio conference, visual representation, speaker content, audience profiles, content analysis, customized feedback, comprehension, interaction, participant monitoring, visualizationThe patent pertains to a system for selective visual representation of speaker content in audio conferences, which involves obtaining profiles for each listener and analyzing spoken content to generate and present tailored visual representations to audience members based on this analysis and their individual profiles. This system analyzes the content voiced by different participants and provides customized visual feedback to different audience members, enhancing the comprehension and interaction within audio conferences. Different visual interpretations of the same audio content are presented to audience members based on their profiles.
2015-06-09 / Speaker association with a visual representation of spoken content / US9053750B2 pdfaudio conference, visual representation, speaker content, monitoring, analysis, display, participant interaction, computing system, conference visualization, content relationshipsThe patented method involves monitoring speaker content from multiple participants in an audio conference and creating a visual representation of this content using a computer with processor and memory. This visual representation is generated for each participant based on the analysis of their spoken content. The result is a display that visually represents the relationships and interactions among speakers and their content within the conference.
2021-05-11 / Gesture control of gaming applications / US11003253B2 pdfimage processing, gaming application, player calibration, computer vision, gesture recognition, interface input devices, pattern recognition, vision system calibration, tracking, image analysisThe patent describes a method and system for calibrating a gaming application by receiving an image of a player and determining a region for interpreting player interactions, aiming to streamline the setup process for computer vision systems that control various applications.
2018-04-03 / Calibrating vision systems / US9933856B2 pdfvision system calibration, human gesture recognition, image processing, display device mapping, computer graphics, gesture-based interface, boundary computation, pattern recognition, data processing, target trackingThe patent describes a system and method for calibrating a vision system by capturing an image of a human gesture that frames a display device, computing the boundary and area of the gesture, and mapping the gesture area to corresponding pixels on the display. This calibration approach eliminates the need for traditional, more cumbersome calibration processes for computer vision systems. The solution is applicable to controlling computers, video games, military vehicles, and medical equipment with gesture-based interfaces.
2016-11-01 / Calibrating vision systems / US9483690B2 pdfvision system calibration, human gesture recognition, image processing, display device, computing boundary, gesture area mapping, pixels, computer graphics, pattern recognition, gesture-based interfaceThe patent describes a vision system calibration method that involves receiving an image of a human gesture framing a display device, computing the boundary and area of the gesture, and subsequently mapping the computed gesture area to corresponding pixels on the display device.
2014-10-14 / Calibrating vision systems / US8861797B2 pdfvision system calibration, human gesture, image processing, boundary computation, gesture area mapping, display device, pattern recognition, computer graphics processing, gesture-based interface, pixel mappingThe patent describes a method and system for calibrating a vision system by interpreting an image of a human gesture that frames a display device, calculating the defined boundary and area of the gesture, and mapping the computed gesture area directly to pixels on the display device to facilitate the calibration process.
2015-12-22 / System and method for dynamic facial features for speaker recognition / US9218815B2 pdfspeaker verification, dynamic image feature, text challenge, biometric identification, facial expression, phonetic content, speech prosody, security, video recording, authentication technologyThis patent introduces a speaker verification system that enhances the robustness of identity authentication by utilizing dynamic image features in conjunction with voice prompts. It records the facial movements and expressions of a user while they speak a uniquely generated text challenge, adding a visual biometric component to traditional voice recognition. This method mitigates security risks associated with solely audio-based verification and circumvents issues like noisy environments or voice playback attacks.
2014-11-25 / System and method for dynamic facial features for speaker recognition / US8897500B2 pdfspeaker verification, dynamic image feature, text challenge, facial movement, phonetic content, speech prosody, facial expression, identity verification, security, biometric identificationThe patent introduces a speaker verification system that enhances security by using unique text challenges paired with dynamic image feature analysis, like facial movements during speech, to verify identity. The dynamic image includes patterns of head, lips, mouth, eyes, and eyebrows movement, considering phonetic content, speech prosody, and facial expressions. This approach addresses traditional speaker verification limitations, such as those due to background noise, and provides robustness against playback attacks.
2010-12-15 / System and method for dynamically and interactively searching media data / 8364673B2 pdfmedia content, interactive search, navigation map, search anchors, anchor cells, dynamic association, document population, query adaptation, visualization, search criteriaThis invention introduces a method and system for interactive media content search, which utilizes a navigation map with cells that are populated with media documents based on user-provided search anchors. The user can dynamically associate the search anchors with anchor cells on the map, allowing for efficient navigation and understanding of the search results in relation to the search criteria. By utilizing this technique, users can adjust their search strategy on-the-fly and explore the media database more effectively, taking into account both the results and the search criteria used.
2012-06-05 / Brief and high-interest video summary generation / US8195038B2 pdfvideo summarization, junk frames, k-means clustering, subshots, shots segmentation, time budget, playback rate, content indicators, hierarchical organization, video analysisThis patent outlines a method for video summarization that includes detecting and adjusting for junk frames, segmenting the video into shots and subshots, using multi-step k-means clustering for grouping subshots, and generating a video summary by evaluating the relative importance of clustered subshots and adhering to a specified time budget. The method further enhances the user experience by rendering the video summary with playback rate information, a marker for the currently playing subshot, and indicators showing segments with similar content. The novelty also lies in the hierarchical approach to video summarization, allowing for a more structured and informative summary.
2018-12-11 / Method and system for analysis of sensory information to estimate audience reaction / US10154295B2 pdfaudience behavior analysis, content viewing correlation, environment influence, display processor, channel selection, correlation value comparison, audience reaction, media processor usage, advertisement placement, Nielsen Television RatingsThe patent describes a system that discerns whether a viewer's channel selection on a display is influenced more by the content being viewed or the viewer's surrounding environment by comparing two correlation values: one representing the connection between the viewer's behavior and the content, and the other representing the connection between the viewer's behavior and environmental factors. If the environment's influence is greater, the system determines this as the cause for a change in channel. This allows for more precise understanding of audience reactions and behaviors with respect to both on-screen content and environmental stimuli.
2017-12-26 / Method and system for analysis of sensory information to estimate audience reaction / US9854288B2 pdfaudience engagement, viewer response, environmental influence, correlation values, channel selection, programming content, real-time behavior, advertisement adjustment, dynamic viewing experience, audience preferencesThe patented system evaluates audience engagement by comparing a viewer's response to content and environmental influences, using correlation values to discern whether channel selections are more influenced by environment than programming content. If the environmental influence exceeds content influence, it assesses whether the viewer's channel choice relates to environmental factors. This allows tailored adjustments to programming or advertisements, providing a dynamic viewing experience based on real-time audience behavior and preferences.
2017-01-31 / Method and apparatus for automated analysis and identification of a person in image and video content / US9558397B2 pdfimage analysis, person identification, metadata, event database, contextual data, image tagging, image retrieval, facial recognition, image categorization, content determinationThe patent introduces a method and system for identifying individuals in an image by utilizing an image analyzer that associates identification metadata with images, based on content and event data from an event database. This metadata includes information about involved persons, locations, and objects, which enhances image tagging and retrieval. Unlike facial recognition, this method relies on a broader set of contextual event data to improve identification accuracy, especially when appearances change.
2017-01-03 / Relational display of images / US9538116B2 pdfimage display, secondary images, relevance, primary image, clustering, content similarity, image retrieval, visual search, storage devices, video framesThe patent describes a method and apparatus for displaying secondary images in correlation with a primary image, where the secondary images' size reflects their relevance to the primary image. The primary image is selectable from multiple images or a video frame, and secondary images are organized into clusters based on similarity of content to the primary image. This system facilitates easier retrieval and review of related images from a large database by visually indicating relevance and content relationship.
2016-06-21 / Method and apparatus for automated analysis and identification of a person in image and video content / US9373024B2 pdfimage analysis, event data, identification metadata, tagging, database, facial recognition, contextual information, metadata generation, image retrieval, appearance changesThe innovation introduces a system that combines image analysis with event data stored in a database to enhance identification and tagging of individuals in images; the process entails generating identification metadata linked to the image and creating tags based on this metadata. The approach addresses limitations of facial recognition by incorporating additional context from event-related information such as related people, locations, and objects from an extendable event database. This allows for more accurate retrieval of images based on user queries, even when appearance changes may affect traditional facial recognition methods.
2016-04-12 / Relational display of images / US9313444B2 pdfimage display, secondary images, primary image, relevance, photo retrieval, clustering, video frame, visual arrangement, content similarity, digital media browsingThe patent describes a method and system for dynamically displaying secondary images in relation to a selected primary image or video frame, where the display size of each secondary image reflects its relevance to the primary image. This relevance-based visual arrangement aids in quick retrieval and viewing of similar content. Clusters of secondary images can be formed to group images with similar content, enhancing the user's browsing experience.
2015-09-15 / Method and system for analysis of sensory information to estimate audience reaction / US9137558B2 pdfaudience engagement, media content analysis, behavioral sensing, environmental characteristic, correlation values, tuning status, media analytics, viewing habits, content influence, environmental influenceThe technical novelty involves a system for determining audience engagement with media content by detecting behavioral patterns of the audience and environmental characteristics, and correlating these to the media being consumed. It differentiates whether the tuning status (staying tuned or switching channels) is due to the content of the media or the surrounding environment of the viewer. If environmental factors have a greater influence on viewing behavior than the content itself, the system identifies the environment as the cause for the tuning status.
2015-08-04 / Relational display of images / US9100616B2 pdfimage display, relevance-based sizing, secondary images, primary image, content similarity, image clusters, navigation, visual content discovery, image dataset, video frame selectionThis invention presents a method and apparatus for dynamically displaying a primary image alongside a cluster of secondary images, where the size of each secondary image reflects its relevance to the primary image. The content similarity between the primary and secondary images is used to form clusters, enabling easier navigation and discovery of related visual content within a large image dataset. A primary image can be sourced from a selection of images or from a video frame, enhancing the user's ability to find and view related images without manual searching.
2015-06-30 / System and method for notification of event of interest during a video conference / US9071728B2 pdfvideo teleconference, event detection, end user notification, computer-implemented method, interactive request, transcription, video conference data, server communication, computer readable medium, speaker commentsThe invention relates to a computer-implemented method that detects events of interest during a video teleconference among participants and notifies an end user about such events, allowing for interactive requests for transcription of the event-related dialogue. The innovation includes a computer-readable medium containing the program that performs these functions. Additionally, after being notified, an end user can utilize their device to contact the server and request transcriptions of comments made by a specific speaker during the teleconference.
2014-12-02 / Relational display of images / US8902361B2 pdfimage display, image data, primary image, secondary images, relevance, size variation, clustering, content similarity, image retrieval, video framesThe patent describes a method and apparatus for organizing and displaying images where secondary images related to a selected primary image are displayed in varying sizes corresponding to their relevance. The secondary images are clustered based on similar content to the primary image, facilitating easier search and retrieval. The system can apply to both individual images and video frames, streamlining the review process of large image collections.
2014-07-29 / Method and apparatus for automated analysis and identification of a person in image and video content / US8792684B2 pdfimage identification, image analyzer, event database, identification metadata, person recognition, content analysis, metadata tagging, image search, facial recognition, event dataThe invention comprises a method, apparatus, and software for enhancing image identification which involves analyzing images to determine content including people, locations, and objects and linking this to event data from a database. It employs identification metadata to tag analyzed images, improving searchability and recognition despite appearance changes. The system is capable of querying an event database to efficiently locate specific images based on user input.
2014-05-06 / Relational display of images / US8717500B1 pdfimage display, secondary images, primary image, relevance, clusters, content similarity, image organization, image search, digital storage, user interfaceThe invention relates to selectively displaying secondary images alongside a primary image, where the secondary images' display sizes are determined by their relevance to the primary image. These images are organized into clusters based on the similarity of their content to that of the primary image. The system streamlines the process of finding and displaying related images from a large collection, such as those from storage devices.
2014-04-15 / System and method for notification of events of interest during a video conference / US8698872B2 pdfcomputer-implemented method, video conference, event detection, end user notification, transcription request, computer-readable medium, program, server communication, data structure, flow chartThe patent describes a computer-implemented method for detecting events of interest during a video conference and providing notifications to end users, with the capability to request transcriptions of a speaker's comments following the event detection. The described system includes a computer-readable medium containing a program that manages these functions, and operates on an end user device that communicates with a server to facilitate this process. The patent includes various flow charts and data structures that depict the operation of the notification and transcription request features in a video conferencing environment.
2018-01-02 / Method and apparatus for distributing content / US9860206B2 pdfmedia content aggregation, origination information, common event identification, viewing perspectives, social network alert, content sharing, smart devices, communication networks, media processing, storage capacityThe patent details a method for aggregating media content from multiple devices based on origination information to identify a common event, determining various viewing perspectives for that event, and subsequently alerting a social network to provide access to these diverse perspectives.
2016-09-06 / Method and apparatus for distributing content / US9438647B2 pdfmedia content aggregation, viewing perspectives, social network notification, origination information, common event, media sharing, content identification, perspective determination, alert system, access facilitationThe technical novelty lies in a system that aggregates media content recorded by various devices at a common event, analyzes the origination information to identify different viewing perspectives, and proactively notifies a social network to facilitate access to these varying perspectives.
2018-02-13 / Augmented multi-tier classifier for multi-modal voice activity detection / US9892745B2 pdfvoice activity detection, multi-tier classifier, multi-modal, audio-visual, feature concatenation, signal-to-noise ratio, classifier output, augmented detection, audio classifier, visual classifierThe patent describes an augmented multi-tier classifier architecture for detecting voice activity using different modalities (e.g., audio and visual signals) simultaneously. It improves audio-visual voice activity detection by concatenating indicators from multiple classifiers with the subject's original features to more accurately determine voice activity. This approach addresses limitations of existing systems that rely on simplistic feature fusion or require knowledge of acoustic signal-to-noise ratio.
2021-11-02 / Dynamic image recognition model updates / US11164294B2 pdfimage recognition, object recognition model, mobile device, electronic device camera, server processing, location information, image capture parameters, image data offloading, server-based computations, network bandwidth optimizationThis invention streamlines the process of offloading resource-intensive image recognition operations from a mobile device to a server, by capturing multiple images under different conditions, identifying objects in these images using an object recognition model, and sending data to the server for processing, which in return provides location information derived from the received data.
2018-10-30 / Dynamic image recognition model updates / US10115185B2 pdfimage recognition, model updating, mobile device, server communication, confidence threshold, dynamic enhancement, resource management, bandwidth reduction, adaptive learning, real-time processingThe patent describes a method for dynamically updating an image recognition model on a mobile device by requesting and receiving updates from a server when the initial recognition results fail to meet a predefined confidence level. The updated model is then used for improved image recognition tasks without offloading the entire process to the server, thus reducing bandwidth and processing resource demands on the mobile device. This enables continuous enhancement of the recognition accuracy and adapts to new image data while operating within resource constraints.
from afar, classroom with multiple robots attaching sticky notes on a whiteboard Hypothetical Patent Process an illustration of truly applied research