Individualized Video Program Generation Using Sensor and Tag Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is a challenge in efficiently capturing and editing video data from multiple sources during large-scale activities, such as sports and events, as it is time-consuming to compile edited video programs for individual participants and spectators, especially with multiple video feeds and participants.
Innovation Solution
A system that stores and edits video data from multiple locations using sensors and video cameras, allowing for continuous recording and triggered recording based on proximity or motion detection, with the ability to generate individualized video programs using sensor and tag data, enabling efficient selection and transfer of video clips for participants and spectators.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple video feeds and sensor data are continuously recorded and stored for all participants, then complete video coverage is achieved, but storage requirements and data management complexity increase significantly
Solution Approach 1:
The patent segments the continuous video stream into discrete clips based on detected events or activities. Instead of managing one large continuous video file, the system divides it into manageable segments that can be independently stored, retrieved, and processed. This segmentation reduces the complexity of data management while maintaining complete coverage of all activities.
Solution Approach 2:
The system extracts only the relevant portions of video data that contain actual activities or events of interest. By using sensor data and activity detection algorithms, the system identifies and extracts meaningful clips from the continuous video feed, discarding unnecessary portions. This extraction approach maintains reliability of coverage for important events while significantly reducing storage requirements and management complexity.
2Adaptability or versatility
If edited video programs are manually compiled for each participant from multiple video feeds, then customized individual videos are created, but the process becomes extremely time-consuming
Solution Approach 1:
The system performs preliminary actions by pre-processing video feeds and sensor data during the event itself. Video clips are automatically detected, segmented, and organized in real-time based on activity recognition. This preliminary organization eliminates the need for manual compilation later, allowing individualized video programs to be quickly generated by simply selecting and assembling pre-processed clips according to participant identity and activity type.
Solution Approach 2:
The system enables self-service video compilation through automated activity detection and clip assembly. Using machine learning models and sensor fusion, the system automatically identifies activities, associates them with the correct participants, and assembles video programs without human intervention. This self-service approach maintains high adaptability for customizing individual videos while dramatically improving productivity by eliminating manual editing processes.
3Manufacturing precision
If video data is stored in high resolution for all participants, then video quality is maintained, but storage space requirements increase
Solution Approach 1:
The patent applies local quality by storing video data at different resolutions based on local requirements. High-resolution video is stored only for segments containing main activities or key participants, while other portions are stored at lower resolutions. This differentiated storage approach maintains video quality where it matters most while significantly reducing overall storage space requirements.
Solution Approach 2:
The system applies partial action by storing complete high-resolution video data only for critical moments or key participants, while using lower-resolution data for less important segments. This selective approach ensures sufficient video quality for the most valuable content while avoiding the excessive storage requirements of storing all video at full resolution.
Data Source
AI summary
Video from multiple locations in a venue, in which multiple individuals are engaged in activities, such as athletic performances or non-sporting activities, is stored and edited, using sensor data from participants and tag data from spectators, to create individualized video programs of athletic performances of individuals. Each camera provides a video feed that is continuously stored. Each video feed is stored in a sequence of data files in computer storage, which data files are created in regular time intervals. Each file is accessible using an identifier of the camera and a time interval. Similarly, data from sensors and tagging devices are continuously received and stored in a database. The database stores, for each sensor, an identifier of each individual detected in the proximity of the sensor and the time at which the individual was detected. Each sensor is associated with one or more cameras. Tag data is similarly stored from tagging devices manipulated by spectators.