Asymmetrical Video Block Boundary Filtering for Discontinuous Error Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing deblocking filtering methods in video coding, such as H.265/HEVC, use symmetrical filters across block boundaries, which can lead to decreased error reduction efficiency when error distributions are discontinuous.
Innovation Solution
An encoder and decoder that determine and apply asymmetrical filter characteristics across block boundaries, where the filter coefficients, weights, or offset values are adjusted based on pixel values, error differences, or quantization parameters to prioritize the reduction of errors in pixels with larger errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If symmetrical filter coefficients are used across block boundaries, then the filtering process is simple and consistent, but error reduction efficiency decreases when error distributions are discontinuous
Solution Approach 1:
The patent applies asymmetry by using different filter coefficients for pixels on opposite sides of the block boundary. Specifically, first filter coefficients are used for pixels in a first direction relative to the boundary, while second filter coefficients are used for pixels in a second direction relative to the boundary. This asymmetric approach allows the filtering to adapt to discontinuous error distributions, thereby improving error reduction efficiency while accepting increased complexity in the filtering process.
Solution Approach 2:
The patent implements local quality by selecting different filter coefficients based on the local characteristics of pixels near the block boundary. The filter coefficients are chosen according to the specific position and direction of each pixel relative to the boundary, allowing the filtering operation to be optimized for local error patterns rather than applying a uniform symmetric filter across the entire boundary region.
2Manufacturing precision
If asymmetrical filter coefficients are applied to all pixels, then error reduction efficiency improves for discontinuous error distributions, but processing complexity and computational load increase
Solution Approach 1:
The patent applies segmentation by dividing the pixels near the block boundary into different groups based on their position and direction relative to the boundary. Pixels are segmented into those in a first direction (using first filter coefficients) and those in a second direction (using second filter coefficients). This segmentation allows the system to apply asymmetrical filtering only where necessary, improving error reduction efficiency while managing computational load through structured organization of the filtering operations.
Data Source
AI summary
An encoder includes processing circuitry and memory coupled to the processing circuitry. Using the memory, the processing circuitry is configured to: change values of luma pixels in a first block and a second block to filter a boundary between the first block and the second block, and change values of chroma pixels in a third block and a fourth block to filter a second boundary between the third block and the fourth block. First luma filter coefficients used for the first block and second luma filter coefficients used for the second block are selected to be asymmetrical with respect to the first boundary based on a prediction mode of the first block and a prediction mode of the second block. First chroma filter coefficients used for the third block and second chroma filter coefficients used for the fourth block are selected to be symmetrical with respect to the second boundary.