Adaptive Deblocking Filter Strength for VVC Block Noise
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In the Versatile Video Coding (VVC) method, the increased maximum block size for prediction leads to increased block noise at block boundaries, and existing deblocking filters operate uniformly without considering the state of block division, resulting in potential loss of texture in fine blocks and inadequate noise reduction in coarse blocks.
Innovation Solution
An information processing device that includes a depth value calculation unit, a filter processing execution determination unit, and a filter strength setting unit to dynamically determine whether and how strongly to apply a deblocking filter based on the depth value of block boundaries, adapting the filter strength according to the block division state to reduce noise while preserving image texture.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If a deblocking filter is applied uniformly to all block boundaries, then block noise is reduced, but texture in fine blocks is lost
Solution Approach 1:
The patent applies different deblocking filter strengths to different block boundaries based on their depth values. Fine blocks (depth value 2) receive weak filtering to preserve texture, while coarse blocks (depth value 0) receive strong filtering to reduce block noise. This local differentiation resolves the contradiction between noise reduction and texture preservation.
Solution Approach 2:
The patent dynamically adjusts the deblocking filter strength according to the depth value of each block boundary. Instead of using a fixed uniform filter strength, the system adapts the filtering intensity based on the hierarchical structure of block division, allowing optimal balance between noise reduction and texture preservation for each specific boundary.
2Productivity
If the maximum block size for prediction is increased, then coding efficiency is improved, but block noise at block boundaries increases
Solution Approach 1:
The patent segments the image into blocks of different sizes and applies differentiated deblocking filter strengths based on block depth. By segmenting the filtering approach according to block hierarchy, the system can maintain large block sizes for coding efficiency while applying appropriate filtering to reduce block noise at each level of the segmentation hierarchy.
3Object-affected harmful factors
If a strong deblocking filter is applied to reduce block noise, then noise reduction is improved, but image quality deteriorates due to texture loss
Solution Approach 1:
The patent applies strong deblocking filter only locally to coarse blocks where block noise is prominent, while applying weak filtering to fine blocks where texture preservation is critical. This localized approach to filter strength maintains overall image quality while effectively reducing block noise in areas where it most occurs.
Data Source
AI summary
A depth value calculation unit (102) calculates, for an image divided into a plurality of blocks, a depth value (Da, Db) for determining presence or absence of processing of a deblocking filter and determining strength of the deblocking filter according to a division state of respective blocks adjacent to a block boundary (Bv). A filter processing execution determination unit (104) determines whether or not to cause the deblocking filter to operate on the block boundary (Bv) on the basis of the depth value (Da, Db). A filter strength setting unit (105) sets the strength of the deblocking filter to operate on the block boundary (Bv) on the basis of the depth value (Da, Db).