Adaptive Filter Grid for Video De-blocking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current de-blocking filters in video stream processing, such as those used in HEVC and JEM with QTBT partitioning, face issues like filtering overlap and reduced parallel processing efficiency due to fixed grid sizes, leading to computational complexity and incomplete edge processing.
Innovation Solution
An adaptive filter grid size is determined based on the number of samples used and modified in de-blocking filtering, allowing for parallel processing of vertical and horizontal edges without overlap, by setting F larger than the sum of samples used and modified in vertical edges and W larger than those in horizontal edges, with F and W being multiples of coding block dimensions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a fixed grid size is used for de-blocking filtering, then the filtering process is simple and consistent, but filtering overlap occurs and parallel processing efficiency is reduced
Solution Approach 1:
The filter grid size is made dynamic rather than fixed. The processor determines the filter grid size adaptively based on the number of samples used and modified in de-blocking filtering of edges. This allows the grid size to adjust to the actual filtering requirements, enabling parallel processing of vertical and horizontal edges without overlap while maintaining processing efficiency.
Solution Approach 2:
The patent changes the parameter of filter grid size from a fixed value to a variable determined by the actual filtering needs. By setting the filter grid size F×W where F and W are larger than the sum of samples used and modified in respective directions, the system optimizes both parallel processing capability and filtering completeness.
2Productivity
If filter grid size is increased to prevent filtering overlap, then parallel processing is enabled, but computational complexity increases
Solution Approach 1:
The filter grid dimensions F and W are set based on the actual number of samples used and modified in de-blocking filtering. This parameter adjustment ensures that the grid is large enough to prevent filtering overlap and enable parallel processing, while not being excessively large to cause unnecessary computational complexity.
3Manufacturing precision
If de-blocking filtering is performed on all edges, then image quality improves, but processing time increases due to sequential processing requirements
Solution Approach 1:
The filtering process is segmented into independent vertical and horizontal edge processing stages. By determining an appropriate filter grid size, the patent enables parallel processing of these segments, allowing all edges to be filtered for complete coverage while reducing overall processing time through parallel execution of independent filtering operations.
Solution Approach 2:
The dynamic filter grid size adaptation allows the system to process all edges thoroughly while enabling parallel execution. The grid size is adjusted to match the actual filtering requirements, ensuring complete edge processing without unnecessary sequential dependencies that would increase processing time.
Data Source
AI summary
A system and method for processing a video stream are provided. When coding or decoding a video stream a de-blocking filtering may be applied for reducing discontinuities caused by use of a block-based video coding. The de-blocking filtering is performed using a filter grid. The size of the filter grid is chosen adaptively based on the processing parameters. The adaptive size of the filtering grid is determined based on the number of samples used and modified in de-blocking filtering of an edge of a coding block. The filtering grid may be used both for vertical and horizontal edge filtering.