Deblocking Filter Context Segmentation for Parallel Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The implementation of deblocking filters in many-core processors for H.264 video compression is inefficient due to data dependence, leading to high startup overhead, synchronization issues, and underutilization of computing power, resulting in image quality discrepancies and reduced performance.
Innovation Solution
A method for deblocking filtering that defines a context area around each macroblock, allowing for independent filtering of macroblocks and utilizing the computing power of multi-core processors, where only a small number of blocks are filtered to maintain high image quality and reduce computing effort, enabling parallel processing and efficient use of available cores.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If deblocking filtering is performed according to standard H.264 with data dependence between blocks, then image quality is maintained, but parallel processing capability is reduced and startup overhead increases
Solution Approach 1:
The patent divides the filtering process into two independent segments: context area filtering and macroblock area filtering. The context area (surrounding blocks) is filtered first and independently, then the macroblock area is filtered using the already-filtered context area as reference. This segmentation eliminates data dependence between adjacent blocks, enabling parallel processing while maintaining image quality.
2Manufacturing precision
If deblocking filtering is performed on all blocks to ensure image quality, then manufacturing precision is improved, but computing effort increases
Solution Approach 1:
The patent applies filtering selectively: the context area surrounding each macroblock is filtered completely, while the macroblock area itself is filtered using a simplified approach that references the already-filtered context area. This partial action approach maintains image quality at block boundaries while reducing computing effort within macroblocks.
3Reliability
If traditional deblocking filtering is used on many-core processors, then synchronization is required between cores, but this increases device complexity and reduces productivity
Solution Approach 1:
The patent performs preliminary filtering of the context area before filtering the macroblock area. By pre-filtering the surrounding blocks and storing their results, each processing core can independently filter its assigned macroblock without requiring real-time synchronization with neighboring cores, thus reducing device complexity and improving productivity.
Data Source
AI summary
A method for deblocking filtering a macroblock on the basis of a coding standard for image information which is stored in macroblocks. A context region comprising the blocks of a first macroblock to be filtered is formable from blocks which are arranged adjacent to one another horizontally in block rows and adjacent to one another vertically in block columns. The blocks of the context region can be filtered on the basis of the coding standard for image information.