Spatially Localized Residual Block Modification for Video Coding Artifacts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards face challenges in transform block decoding and encoding, particularly in reducing visual artifacts like 'ringing' due to strong quantization, and struggle to efficiently encode areas with both smooth gradients and local high frequencies, leading to inefficient coding or artifacts.
Innovation Solution
The method involves inverse transforming transform blocks to obtain residual blocks and modifying specific sub-portions of the reconstructed prediction error values based on other residual blocks or transform skip blocks, allowing for spatially localized improvements that compensate for artifacts and enhance encoding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a large transform block size is selected to encode smooth gradients efficiently, then coding efficiency is improved, but ringing artifacts appear in local high frequency parts
Solution Approach 1:
The transform block is divided into multiple sub-blocks, allowing different transform sizes to be applied to different regions. This segmentation enables the encoder to use large transform blocks for smooth gradient areas (improving coding efficiency) while using smaller transform blocks for local high frequency parts (avoiding ringing artifacts).
Solution Approach 2:
Different transform block sizes are applied to different spatial regions based on local content characteristics. Smooth gradient regions use larger transform blocks for efficient coding, while regions with local high frequencies use smaller transform blocks to prevent artifacts, achieving local optimization of both coding efficiency and artifact reduction.
2Object-affected harmful factors
If a small transform block size is selected to encode local high frequencies, then artifact reduction is improved, but coding efficiency deteriorates due to repeated encoding of smooth gradients
Solution Approach 1:
The transform block is divided into multiple sub-blocks, allowing different transform sizes to be applied to different regions. This segmentation enables the encoder to use small transform blocks only where necessary (local high frequency parts for artifact reduction) while using large transform blocks for smooth gradient areas (maintaining coding efficiency).
Solution Approach 2:
Different transform block sizes are applied to different spatial regions based on local content characteristics. Regions requiring artifact reduction use smaller transform blocks, while smooth gradient regions use larger transform blocks for efficient coding, achieving local optimization of both artifact reduction and coding efficiency.
3Productivity
If transform coding is applied to remove redundancy in prediction errors, then compression efficiency is improved, but visual artifacts appear due to strong quantization
Solution Approach 1:
The transform block is divided into multiple sub-blocks, allowing different quantization strengths to be applied to different regions. This enables strong quantization (improving compression efficiency) in regions where it is acceptable, while using weaker quantization or transform skip in regions where visual quality is critical, thereby reducing overall visual artifacts while maintaining compression efficiency.
Solution Approach 2:
Different quantization parameters are applied to different spatial regions based on local content characteristics and importance. Less important regions undergo strong quantization for compression efficiency, while important regions use weaker quantization to preserve visual quality and avoid artifacts, achieving local optimization of both compression and quality.
Data Source
AI summary
Reconstructed prediction errors in a sub-portion (12) of a residual block (11) obtained by inverse transforming a transform block (10) are modified in order to spatially improve localized portions of the residual block (11), e.g. to compensate for visual artifacts from transform coding. The modification affects reconstructed prediction errors in the sub-portion (12) of the residual block (11) but not reconstructed prediction errors in a remaining portion of the residual block (11).