Server-Side Video Region Segmentation for Bandwidth Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Video conferencing systems face issues with bandwidth constraints and distracting backgrounds, leading to degraded video quality, as participants' network conditions and backgrounds can significantly impact the receiving end, making it uncomfortable to request adjustments from the transmitting participant.
Innovation Solution
A method to adjust video settings, such as blurring or modifying backgrounds, at the server level before transmission, allowing receiving participants to improve video quality without involving the sender, based on bitrate, display size, or aesthetic concerns, with automatic or user-initiated requests.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If video is transmitted at high resolution, then video quality at the capturing end is improved, but transmission bandwidth requirements increase and video quality at the receiving end deteriorates
Solution Approach 1:
The patent applies local quality by transmitting video at different resolutions for different regions: high resolution for the speaker area and low resolution for the background area. This allows the system to focus bandwidth on the most important visual element (the speaker) while reducing bandwidth consumption overall, resolving the contradiction between video quality and bandwidth usage.
Solution Approach 2:
The patent segments the video feed into distinct regions (speaker area and background area) and processes each region differently. By separating the video stream into multiple quality versions for different regions, the system can optimize bandwidth allocation and maintain acceptable video quality while reducing overall transmission bandwidth requirements.
2Loss of information
If complex backgrounds are transmitted, then visual fidelity is improved, but bandwidth requirements exceed available capacity causing degraded video quality
Solution Approach 1:
The patent extracts and removes the background portion from the video feed, transmitting only the speaker area at high quality. This extraction approach eliminates the bandwidth burden of complex backgrounds while preserving the visual fidelity of the important speaker region, directly resolving the contradiction between visual fidelity and bandwidth availability.
Solution Approach 2:
The patent changes the quality parameter for different video regions dynamically. The speaker area is transmitted at high quality parameters while the background is either removed or transmitted at low quality parameters. This parameter differentiation allows the system to maintain visual fidelity where needed while adapting to bandwidth constraints.
3Ease of operation
If receiving participant requests background adjustments, then visual experience is improved, but involving the transmitting participant becomes disruptive and uncomfortable
Solution Approach 1:
The patent introduces an intermediary mechanism (automatic background detection and processing system) that mediates between the receiving participant's visual experience needs and the transmitting participant's comfort. The system automatically detects speaker regions and adjusts backgrounds without requiring direct communication or involvement between participants, eliminating social discomfort while improving visual experience.
Solution Approach 2:
The patent implements self-service by enabling the receiving participant's device to automatically adjust and optimize background display without needing to contact or coordinate with the transmitting participant. The system autonomously processes video feeds to enhance visual experience, removing the harmful social interaction element while maintaining ease of operation.
Data Source
AI summary
Systems and methods are provided herein for altering video feeds of a first device in a video conference according to a request of a second device in the video conference. The present disclosure identifies a first video captured by a first device for a video conference involving the first device and a second device, receives a user request for modification of video transmitted to a second device regarding the first video, adjusts the first video according to the user request to create a second video, wherein the second video is visibly distinct from the first video, and transmits the created second video for display at the second device.