Parallel Deblocking Filter for Image Data Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Deblock filtering processing in image encoding, such as in MPEG4/AVC methods, requires significant calculation and time due to dependency relations between macro blocks, making parallel processing difficult and thus challenging to shorten processing time.
Innovation Solution
A data processing apparatus and image processing apparatus that perform predetermined processing on element data of first and second blocks in parallel by using processing results from each other, with specific processing means for horizontal and vertical filtering that do not rely on immediate pixel data from adjacent blocks, allowing for independent processing and result combination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If deblock filtering processing is performed sequentially using processing results from adjacent blocks, then processing accuracy is maintained, but processing time becomes excessively long
Solution Approach 1:
The patent divides the image data into multiple blocks and segments the filtering processing into independent units. Each block's filtering can be performed independently by selecting appropriate reference blocks, enabling parallel processing while maintaining accuracy. The segmentation allows different blocks to be processed simultaneously without interfering with each other's accuracy requirements.
Solution Approach 2:
The patent performs preliminary identification of valid reference blocks before actual filtering processing. By pre-determining which blocks contain valid reference data and organizing this information in advance, the system eliminates waiting time during filtering operations and enables parallel processing to proceed without delays caused by dependency checks.
2Productivity
If parallel processing is implemented for deblock filtering, then processing time is reduced, but dependency relations between blocks cause processing difficulties
Solution Approach 1:
The patent introduces a reference block management mechanism that acts as an intermediary between blocks. This mechanism tracks which blocks have been processed and which contain valid reference data, allowing parallel processing units to independently determine their dependencies without direct communication. The intermediary structure simplifies the complexity of managing inter-block dependencies in parallel processing.
Solution Approach 2:
The patent changes the processing parameters by allowing flexible selection of reference blocks based on spatial relationships and processing status. Instead of fixed sequential dependencies, the system dynamically determines valid reference blocks using parameters such as block positions, processing completion flags, and validity markers, enabling parallel execution while maintaining correct filtering accuracy.
3Manufacturing precision
If filtering processing uses processing results from adjacent blocks, then filtering quality is improved, but processing cannot be performed in parallel
Solution Approach 1:
The patent segments the image into independent processing blocks and identifies specific reference block relationships for each segment. By determining valid reference blocks in advance and organizing them by spatial position, the system enables each block to perform high-quality filtering using its designated references while allowing simultaneous processing of multiple blocks, thus achieving both quality and parallelism.
Solution Approach 2:
The patent performs preliminary analysis to identify and store valid reference block information before filtering processing begins. This pre-computation of reference relationships allows filtering operations to proceed in parallel without needing to check dependencies during execution, maintaining filtering quality while enabling full parallel processing capability across all blocks.
Data Source
AI summary
A data processing apparatus capable of performing processing in units of block data by using processing results of other block data in a short time is provided, wherein a deblocking filter performs horizontal filtering processing and vertical filtering processing on first block data and second block data obtained by dividing restructured picture data into two in parallel by adjusting their dependency relation.