Personalized Live Video Feed Generation via Subject Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current live video feed streaming systems fail to provide personalized views based on user preferences, leading to decreased user interest and increased computational delays, and require extensive computing infrastructure, resulting in higher costs without offering personalized benefits.
Innovation Solution
A method and system that utilizes a view generator to determine personalized inputs from users, correlate real-time subject locations with event maps, and select and transmit tailored video feeds, allowing users to focus on specific subjects or incidents, thereby enhancing the viewing experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If image processing systems are used to provide different views of live video feeds, then users can access multiple perspectives, but additional time delays occur due to complex computation and post processing
Solution Approach 1:
The system pre-processes video feeds from multiple cameras during the event, preparing personalized views in advance based on predicted user interests. This preliminary action allows the system to deliver personalized content without real-time processing delays, as the content is ready when users request it.
Solution Approach 2:
The system segments video content into multiple personalized streams based on user preferences, interests, and behavior patterns. Each user receives a customized feed composed of relevant segments from multiple camera sources, eliminating the need to process all feeds for all users and reducing overall computational delay.
2Speed
If extensive computing infrastructure is deployed to perform fast and accurate image processing, then processing speed improves, but system complexity and costs increase
Solution Approach 1:
The system extracts only the essential processing operations needed for personalization, separating critical real-time tasks from optional enhancements. By focusing computational resources on key functions like subject identification and feed selection, the system achieves fast processing without requiring extensive infrastructure.
Solution Approach 2:
Instead of processing all video feeds in real-time for every user, the system creates simplified copies or representations of video content that can be quickly assembled into personalized feeds. This approach reduces processing complexity while maintaining the ability to deliver customized views rapidly.
3Adaptability or versatility
If traditional streaming systems broadcast live video feeds, then all users receive the same content, but user preferences and personalized viewing experiences are not considered
Solution Approach 1:
The system applies local quality by customizing video feed content according to each user's specific preferences, interests, and behavior patterns. Different users receive different combinations of camera feeds and content types, creating personalized viewing experiences that match individual tastes and increase engagement.
4Area of stationary object
If multiple cameras are used to provide different views, then users can see various perspectives, but the system cannot identify subjects at long distances
Solution Approach 1:
The system merges data from multiple camera feeds and auxiliary data sources (such as tracking systems, metadata, and historical information) to identify and track subjects across wide areas. By combining information from various sources, the system overcomes the limitation of individual cameras and achieves accurate subject identification even at long distances.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present disclosure discloses a method and a system for generating personalized views of live video feeds of an event. The method comprises determining one or more personalized inputs associated with each user. Further, the method comprises receiving a live video feed of an event from each of a plurality of imaging units. Then, the method generates a map of an area within which the event occurs. The method further comprises, receiving real-time location of at least one subject participating in the event from a tracking unit, correlating the real-time location of the subject with the mapped area, selecting, for each user, at least one video feed among the live video feeds for generating personalized views of the live video feeds and transmitting the at least one selected video feed to respective users through a web server for providing personalized views of the live video feeds of the event.