Deblocking Filter Application at Sub-Block Boundaries in VVC
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In image processing, block distortion occurs at sub-block boundaries in Intra Sub-Partitions (ISP) due to the lack of deblocking filter application, particularly for color difference components in Versatile Video Coding (VVC), where the existing deblocking filter is only applied to luminance components and not effectively to sub-block boundaries.
Innovation Solution
An image processing device and method that decodes a bitstream to generate a decoded image with a hierarchical block structure, applying a deblocking filter to pixels orthogonal to sub-block boundaries, thereby suppressing block distortion by extending deblocking filter application to sub-block boundaries in both luminance and color difference components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If blocks are divided into sub-blocks in Intra Sub-Partitions (ISP) for improved coding efficiency, then coding efficiency is improved, but block distortion occurs at sub-block boundaries because the deblocking filter is not applied to these boundaries
Solution Approach 1:
The transform block is divided into multiple sub-blocks for independent orthogonal transform processing, which improves coding efficiency by adapting to local frequency characteristics. The deblocking filter is then applied to the boundaries between these segmented sub-blocks to suppress the distortion caused by segmentation.
Solution Approach 2:
The deblocking filter is selectively applied to sub-block boundaries based on local conditions. The filter strength and application are determined by analyzing pixel differences and gradient values at specific locations adjacent to sub-block boundaries, allowing targeted suppression of block distortion where it occurs most severely.
2Object-affected harmful factors
If the deblocking filter is applied to sub-block boundaries to suppress block distortion, then block distortion is suppressed, but processing complexity increases
Solution Approach 1:
Instead of applying the deblocking filter uniformly to all boundaries, the patent applies it selectively to sub-block boundaries based on specific conditions. The filter is applied with varying strength depending on the local pixel characteristics, performing partial action only where needed to suppress block distortion effectively.
Solution Approach 2:
The deblocking filter parameters (strength, application range) are dynamically changed based on local image characteristics. The filter strength is adjusted according to the absolute difference of pixel values and gradient calculations at sub-block boundaries, allowing adaptive suppression of block distortion while managing processing complexity.
Data Source
AI summary
In accordance with a block structure having a hierarchical structure in which a transform block, a prediction block, and a coding block are unified, a bitstream is decoded and a decoded image is generated. A deblocking filter is applied to pixels of a line orthogonal to a sub-block boundary that is a boundary between sub-blocks obtained by dividing the transform block of the decoded image. The present technology can be applied to, for example, a case where encoding and decoding of an image are performed.