Visual Conditioning for AR Video Conferencing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In telepresence video conferencing, the integration of multiple video objects with distinct backgrounds results in a distracting and disorienting visual experience due to segmentation errors, which existing image segmentation techniques fail to adequately address, particularly on mobile devices with limited computational resources.
Innovation Solution
A method for visually conditioning augmented-reality-assisted video conferencing that separates foreground and background, designs directional textures to hide segmentation artifacts, and combines them with harmonized backgrounds to create a seamless composite image, utilizing principles from graphic design and contour hatching techniques, and leveraging depth sensors for improved segmentation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If available image/video segmentation techniques are used to separate human subjects from backgrounds, then visual separation is achieved, but video conference rendering quality deteriorates due to segmentation errors
Solution Approach 1:
The patent applies color and texture modifications to the background region to mask segmentation errors. By changing the visual properties of the background (applying textures, adjusting colors, adding patterns), the segmentation inaccuracies become less noticeable, improving perceived quality without requiring more complex segmentation algorithms
Solution Approach 2:
The patent introduces an intermediary processing step between segmentation and final rendering. A background modification module acts as a mediator that processes the segmented background to hide artifacts, allowing simple segmentation to produce high-quality final images through intermediate enhancement
2Adaptability or versatility
If multiple video objects with distinct backgrounds are transmitted and displayed together, then telepresence functionality is achieved, but visual distraction and disorientation increase
Solution Approach 1:
The patent applies different visual conditioning treatments to different regions of the composite video. Each participant's background is independently modified with appropriate textures and patterns, allowing localized optimization to reduce visual distraction while maintaining the multi-participant telepresence functionality
Solution Approach 2:
The patent modifies background colors and applies visual textures to create harmonized backgrounds across multiple video feeds. This color and texture coordination reduces visual disorientation caused by multiple distinct backgrounds while preserving the ability to display multiple participants simultaneously
3Device complexity
If simple background blurring is applied to modify the background, then processing complexity is reduced, but segmentation artifacts remain visible
Solution Approach 1:
Instead of simple blurring, the patent applies color modifications, texture overlays, and pattern generation to the background. These visual transformations effectively mask segmentation artifacts by changing the background's visual characteristics, achieving better artifact hiding with comparable processing complexity
Solution Approach 2:
The patent creates a composite background by combining the original background with added textures, patterns, and visual elements. This composite approach masks segmentation errors more effectively than simple blurring while maintaining processing efficiency
Data Source
AI summary
An apparatus comprising a processor configured to receive a video teleconferencing image, wherein the video teleconferencing image has a background and at least one foreground, wherein the foreground has a boundary, determine the boundary of the foreground, overlay the boundary of the foreground with a drawing area, wherein the drawing area obscures the boundary, and replace the background of the image with a new background.