Asymmetric Deblocking Filter for Video Block-Edge Artifact Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies suffer from visible edge artifacts at block edges due to block-based coding, necessitating improved deblocking filters to enhance image quality.
Innovation Solution
A deblocking filter apparatus and method that utilizes an asymmetric filter to modify sample values on either side of a block edge, adjusting the number of modified samples based on available line buffer size to optimize filtering efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If a conventional deblocking filter is applied to remove edge artifacts at block edges, then picture quality is improved, but computational complexity increases and filtering efficiency decreases when line buffer resources are limited
Solution Approach 1:
The patent dynamically changes the filter strength parameter based on the available line buffer size. When line buffer is sufficient, strong filtering is applied to maximize artifact removal. When line buffer is limited, filter strength is reduced to maintain processing efficiency. This parameter adaptation resolves the contradiction by adjusting filtering intensity to available resources.
Solution Approach 2:
The deblocking filter transitions from a static, uniform filtering approach to a dynamic approach where filtering decisions are made in real-time based on line buffer availability. The filter adapts its behavior during processing, switching between different filtering modes depending on resource constraints, thereby balancing quality and complexity dynamically.
2Object-affected harmful factors
If strong deblocking filtering is applied to maximize artifact removal, then edge artifact visibility is reduced, but processing time increases and filtering efficiency decreases
Solution Approach 1:
The patent adjusts the filtering parameter based on the strength of block artifacts detected. When artifacts are strong, higher filtering strength is applied. When artifacts are weak, filtering strength is reduced to maintain efficiency. This adaptive parameter change resolves the contradiction between artifact removal and processing speed.
Solution Approach 2:
The patent applies filtering selectively rather than uniformly across all block edges. Filtering is applied with appropriate strength only where needed based on artifact detection, avoiding unnecessary processing in regions with minimal artifacts, thereby maintaining efficiency while effectively removing visible artifacts.
3Manufacturing precision
If the number of modified samples is increased to improve deblocking performance, then block artifact reduction is enhanced, but available line buffer resources are exceeded
Solution Approach 1:
The patent dynamically adjusts the number of samples to be filtered (filter span) based on the available line buffer size. When line buffer is large, more samples are processed to improve deblocking performance. When line buffer is small, the number of samples is reduced to fit within available memory resources. This resolves the contradiction by adapting the processing scope to resource availability.
Solution Approach 2:
The filter dynamically adjusts its operational scope (number of samples processed) based on real-time assessment of line buffer availability. This dynamic adaptation allows the system to maximize deblocking performance when resources permit while maintaining operational feasibility when resources are constrained.
4Measurement precision
If asymmetric filtering is applied to handle different block sizes, then filtering accuracy is improved, but device complexity increases
Solution Approach 1:
The patent applies asymmetric filtering where the number of samples filtered on each side of the block edge is adjusted based on the relative sizes of adjacent blocks. For example, when one block is larger than its neighbor, more samples are filtered on the larger block's side. This asymmetric approach improves accuracy for varying block sizes while maintaining a relatively simple filter structure.
Solution Approach 2:
The filter applies different filtering characteristics to different local regions based on block size variations. Each block edge is processed with filtering parameters tailored to the specific local configuration of adjacent blocks, improving local accuracy without requiring a completely complex global filter structure.
Data Source
AI summary
A method and image processing device are provided, including a deblocking filter. The deblocking filter modifies values of at most MA samples of the first image block as first filter output values, the at most MA samples being located at a column of the first image block that is perpendicular to and adjacent to the horizontal block edge; and modifies values of at most MB samples of the second image block as second filter output values, the at most MB samples being located at a column of the second image block that is perpendicular to and adjacent to the horizontal block edge. At most a number MA of sample values of the first image block adjacent to the block edge are modified and at most a number MB of sample values of the second image block adjacent to the block edge are modified, wherein MA<MB.