Block Size Restriction for Illumination Compensation in Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently processing video data due to high processing complexities and bandwidth requirements, particularly when applying illumination compensation for certain block sizes, which can lead to increased memory and computational costs.
Innovation Solution
The proposed solution involves selectively disabling illumination compensation for specific block sizes, such as 4×4 blocks in the Versatile Video Coding (VVC) standard, and blocks larger than the pipeline structure, to reduce processing complexity and avoid additional memory bandwidth issues by determining whether bi-directional inter-prediction is restricted and adjusting the illumination compensation flag accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If illumination compensation is applied to all blocks, then video quality is improved, but processing complexity and memory bandwidth increase
Solution Approach 1:
The patent applies illumination compensation selectively based on block size characteristics. Instead of uniformly applying the technique to all blocks, it identifies specific block size ranges (e.g., 4×4 blocks in VVC, or blocks exceeding pipeline structure dimensions) where the technique provides benefit while avoiding application to blocks where it would increase complexity without sufficient gain. This localized application strategy resolves the contradiction by adapting the processing intensity to the specific characteristics of each block.
2Manufacturing precision
If illumination compensation is applied to small blocks (e.g., 4×4), then local illumination variations are compensated, but memory bandwidth and computational cost increase
Solution Approach 1:
The patent changes the parameter of block size threshold to control when illumination compensation is applied. By establishing specific size thresholds (such as restricting 4×4 blocks or blocks larger than pipeline structure), the system adjusts the application criteria based on block dimensions. This parameter-based control resolves the contradiction by preventing illumination compensation from being applied to block sizes where the memory bandwidth cost outweighs the quality benefit.
3Device complexity
If bi-directional inter-prediction is restricted for certain block sizes, then processing complexity is reduced, but illumination compensation may still be applied inconsistently
Solution Approach 1:
The patent extends the block size restriction principle from bi-directional inter-prediction to illumination compensation as well. By applying the same block size criteria to both techniques, the system ensures that when bi-directional inter-prediction is restricted for certain block sizes, illumination compensation is also restricted for the same block sizes. This universal application of the restriction principle resolves the contradiction by maintaining consistency across different prediction techniques and avoiding the inconsistency that would arise from applying different rules to different techniques.
Data Source
AI summary
Systems, methods, and computer-readable storage media for restricting block sizes for illumination compensation are described. An example method can include obtaining a block of a picture of video data; determining a size of the block; determining whether the size of the block is equal to or less than a first block size associated with a block size restriction that restricts bi-directional inter-prediction for blocks having a respective size that is equal to or less than the first block size associated with the block size restriction or whether the size of the block is greater than a second block size associated with a video coding pipeline structure; and disabling illumination compensation for the block based on a determination that the size of the block is equal to or less than the first block size or greater than the second block size.