Video Interruption Removal via Deep Learning Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Video conferencing systems often experience interruptions due to unexpected objects or noise in the user's surroundings, disrupting the conference and affecting other participants.
Innovation Solution
A video conferencing system that utilizes a deep learning model to detect and remove interruption objects from the video stream in real-time, using a processor coupled with an image capturing device and storage, allowing for immediate correction without affecting other participants.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional video conferencing systems are used without interruption detection, then the system complexity is low, but the video conference is frequently interrupted by unexpected objects or noise
Solution Approach 1:
The system performs preliminary detection of the video frame using a deep learning model before the interruption occurs, identifying potential interruption objects in advance and removing them proactively from the video stream to prevent conference disruptions
Solution Approach 2:
A deep learning model is introduced as an intermediary component between the image capturing device and the video conference output, serving as a mediator that detects and removes interruption objects while allowing the main video conference functionality to continue uninterrupted
2Reliability
If deep learning model is used to detect and remove interruption objects, then the video conference continuity is improved, but the processing time and computational resources increase
Solution Approach 1:
The deep learning model processes video frames at periodic intervals rather than continuously analyzing every frame, detecting interruption objects at specific time points and removing them to maintain conference continuity while reducing overall processing time
Solution Approach 2:
The system applies partial action by focusing the deep learning model's detection capability only on specific regions or types of objects that are likely to cause interruptions, rather than analyzing the entire video frame in detail, thus reducing processing time while maintaining effectiveness
Data Source
AI summary
A video conferencing system and a method of removing an interruption thereof are provided. The method includes the following steps. A video conference is activated and a video stream is obtained through an image capturing device. A deep learning model is used to detect at least one first image object in a first video frame of the video stream. Whether the at least one first image object is an interruption object is determined. The at least one first image object is removed from the first video frame in response to the at least one first image object being determined to be the interruption object.