Video Background Cleanup Neural Network
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video conferencing software lacks the ability to selectively alter portions of a background view during a video conference without using a full virtual background, failing to address the need to conceal unwanted items or enhance bland backgrounds.
Innovation Solution
The implementation of background cleanup systems that store a reference image of a physical background, allowing for the identification and removal of extraneous items during a video conference, and optionally using replacement imagery prediction to enhance the background.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If conventional video conferencing software uses a full virtual background, then unwanted items in the background are concealed, but the user cannot share their real background and loses authenticity
Solution Approach 1:
The patent segments the background processing into two distinct modes: full virtual background replacement and selective background cleanup. This allows users to choose between complete background substitution or partial background modification, resolving the contradiction between concealing unwanted items and maintaining background authenticity.
Solution Approach 2:
The patent applies local quality by enabling selective modification of specific portions of the background while preserving the rest. The background cleanup feature allows users to remove only unwanted items (such as clutter or personal belongings) from specific regions of the background, rather than replacing the entire background, thus maintaining authenticity while eliminating harmful visual elements.
2Object-affected harmful factors
If conventional video conferencing software uses a full virtual background, then unwanted items are concealed, but the setup and configuration become more complex
Solution Approach 1:
The patent extracts the complexity of full virtual background replacement by offering a simplified alternative - background cleanup. This feature removes only the unwanted items from the background without requiring users to configure entire virtual backgrounds, thereby reducing setup complexity while still achieving the goal of concealing unwanted elements.
Solution Approach 2:
The background cleanup feature acts as a lightweight, on-demand solution compared to the heavy, permanent setup of full virtual backgrounds. Users can apply background cleanup selectively and temporarily for specific conferencing needs without committing to complex virtual background configurations, making the system more adaptable and easier to use.
3Object-affected harmful factors
If conventional video conferencing software uses a full virtual background, then unwanted items are concealed, but processing resources and energy consumption increase
Solution Approach 1:
The patent applies partial action by implementing background cleanup that removes only unwanted items from the background rather than processing and replacing the entire background. This selective approach reduces the computational load and energy consumption compared to full virtual background replacement, while still achieving the desired effect of concealing unwanted elements.
4Object-affected harmful factors
If conventional video conferencing software uses a full virtual background, then unwanted items are concealed, but the background appears less natural and professional
Solution Approach 1:
The background cleanup feature applies local quality by selectively removing unwanted items from specific regions of the background while preserving the natural appearance of the rest of the background. This creates a more authentic and professional look compared to full virtual background replacement, as the user's real background remains visible with only problematic elements removed.
Data Source
AI summary
A computer identifies, using software-based image processing applied to camera-generated visual data from a client device, foreground imagery representing a participant and background imagery representing content of the camera-generated visual data other than the foreground imagery. The computer determines, using a neural network, an identity of an extraneous item comprising a portion of the background imagery. The computer determines, using the neural network based on the identity of the extraneous item, to replace the extraneous item in the camera-generated visual data. The computer predicts, using replacement imagery prediction software, replacement imagery to replace the extraneous item. The computer generates a composite image comprising the foreground imagery of the camera-generated visual data and the background imagery with the extraneous item replaced by the replacement imagery.


