Residual Video Coding With Selective Background Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding methods encode the residual data of the whole picture, leading to high compressed average code rates, affecting viewing experience under limited bandwidth, and allocate the same code rate to both high significance object areas and low significance background areas, wasting resources.
Innovation Solution
Dynamically determine whether to encode the residual data of background areas in inter predicted pictures, using object masks and dilation processing to selectively encode only the object area residuals, and adaptively adjust code rates based on impact factors and similarity calculations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If whole picture residual data is encoded uniformly, then encoding simplicity is maintained, but code rate resources are wasted on background areas
Solution Approach 1:
The picture is segmented into object area and background area based on motion vector differences. Object area pixels have motion vector differences within a threshold, while background area pixels exceed the threshold. This segmentation enables differential encoding where only object area residual data is encoded, eliminating waste of code rate resources on background areas while maintaining encoding efficiency.
Solution Approach 2:
Different encoding strategies are applied to different regions: object area uses full residual encoding to maintain quality, while background area uses simplified or skipped encoding to save resources. This local quality approach ensures that code rate resources are allocated according to the actual importance and complexity of each region.
2Manufacturing precision
If object area residual is prioritized, then subjective video quality is improved, but encoding complexity increases
Solution Approach 1:
Motion vectors are calculated and compared in advance to identify object and background areas before residual encoding. This preliminary classification based on motion vector differences allows the encoder to prepare region masks and determine which areas require full encoding, thereby improving video quality without adding significant complexity during the actual residual encoding stage.
3Reliability
If background residual compensation is performed, then encoding completeness is maintained, but bandwidth consumption increases
Solution Approach 1:
The background area residual encoding is extracted and removed from the overall encoding process. By identifying background pixels through motion vector difference thresholding and creating a background mask, the encoder excludes background areas from residual encoding, thereby reducing bandwidth consumption while maintaining encoding completeness for the important object areas.
Data Source
AI summary
A residual coding method is provided. The method comprises: when the current frame is an inter-frame prediction frame, determining whether a residual of a background region in the current frame needs to be compensated for; and when the residual of the background region in the current frame needs to be compensated for, coding first residual data, and when the residual of the background region in the current frame does not need to be compensated for, coding second residual data, wherein the first residual data is residual data of an entire image of the current frame, and the second residual data is residual data of a target region in the current frame. The present disclosure further provides a video coding method using the residual coding method, and corresponding devices and a storage medium.