Recursive Intra Region Transform Partitioning for Chroma Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently compressing chroma blocks with larger block sizes, leading to accuracy losses due to inflexible transform partitioning, which can be addressed by employing different transform partitioning for chroma and luma blocks within a coding region.
Innovation Solution
Implementing distinct transform partitioning for chroma and luma blocks within a coding region, allowing chroma blocks to have smaller partitions while restricting luma blocks to reduce overhead, thereby improving image quality and minimizing data loss.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If chroma blocks use larger block sizes, then compression efficiency is improved, but image quality accuracy deteriorates
Solution Approach 1:
The patent applies different transform partitioning strategies to chroma and luma blocks based on their specific characteristics. Chroma blocks receive more aggressive transform partitioning (smaller partitions) to compensate for their larger block sizes and maintain accuracy, while luma blocks use standard partitioning. This local differentiation resolves the contradiction by tailoring the transform strategy to each block type's specific needs.
Solution Approach 2:
The patent segments the transform partitioning process into different modes for chroma and luma blocks. By separating the transform partitioning into distinct paths based on block type, the system can optimize chroma blocks with finer partitions while maintaining luma block efficiency, thus resolving the accuracy-efficiency tradeoff.
2Measurement precision
If transform partitioning flexibility is increased for chroma blocks, then image quality improves, but data overhead increases
Solution Approach 1:
The patent applies selective transform partitioning where chroma blocks receive enhanced partitioning flexibility only when needed (e.g., when chroma block size exceeds a threshold). This localized application of flexibility improves image quality for chroma blocks without unnecessarily increasing overhead for all blocks, resolving the quality-overhead contradiction.
Solution Approach 2:
The patent applies transform partitioning excessively (more aggressively) to chroma blocks compared to luma blocks, recognizing that chroma blocks benefit more from fine partitioning due to their larger sizes and lower frequency content. This partial excessive action on chroma blocks achieves quality improvement with acceptable overhead cost.
3Productivity
If chroma blocks use larger block sizes, then compression efficiency improves, but transform partitioning accuracy deteriorates
Solution Approach 1:
The patent segments the transform partitioning into separate handling paths for chroma and luma blocks. Chroma blocks are subjected to additional transform partitioning steps that segment them into smaller sub-blocks, which improves transform partitioning accuracy by better matching the frequency content characteristics of chroma data.
Solution Approach 2:
The patent changes the transform partitioning parameters (block size, partition depth, partition mode) specifically for chroma blocks based on their larger size and frequency characteristics. This parameter adaptation resolves the accuracy deterioration by optimizing transform parameters to match chroma block properties.
Data Source
AI summary
An example method of video coding includes receiving a video bitstream including a plurality of frames; identifying a coding region in a frame of the plurality of frames, the coding region including two or more blocks of the plurality of blocks, where each block in the coding region is encoded in an intra prediction mode. The method includes obtaining a set of luma transform blocks according to a first transform partitioning for luma blocks in the coding region and obtaining a set of chroma transform blocks according to a second transform partitioning for chroma blocks in the coding region, where the second transform partitioning is different than the first transform partitioning. The method further includes reconstructing the two or more blocks of the coding region using respective prediction blocks, the set of luma transform blocks, and the set of chroma transform blocks.