Adaptive Loop Filter Preprocessing With Residual Gradients
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies require increased hardware complexity due to the need for buffering residual and reconstructed samples for adaptive loop filtering (ALF), which is performed after deblocking and SAO filtering.
Innovation Solution
Process residual data to determine processed residual data, which requires less memory and is used for ALF, reducing the need for extensive buffering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If residual samples and reconstructed samples are buffered for ALF, then ALF can be performed, but hardware complexity and memory requirements increase
Solution Approach 1:
The patent extracts only the essential information from residual samples by computing gradients (differences between adjacent samples) and storing only these gradient values instead of the complete residual samples. This extraction principle reduces the data volume requiring buffering while preserving the information needed for ALF filter determination, thereby resolving the contradiction between maintaining ALF performance and reducing hardware complexity
Solution Approach 2:
The patent segments the residual block into multiple smaller gradient calculation units, where gradients are computed for specific sample positions rather than buffering entire residual blocks. This segmentation approach divides the large memory requirement into smaller, manageable gradient values, reducing overall memory requirements while enabling ALF operations
2Measurement precision
If residual samples are buffered for ALF, then filter selection can be improved, but memory requirements increase
Solution Approach 1:
The patent changes the parameter representation from storing actual residual sample values to storing gradient values (differences between adjacent samples). This parameter transformation reduces the quantity of data requiring memory storage while preserving the essential information needed for accurate filter selection during ALF, thus resolving the contradiction between measurement precision and quantity of substance
Data Source
AI summary
A device for decoding video data can be configured to determine a prediction block for a current block of the video data; determine a residual block for the current block of the video data; process residual data of the residual block to determine processed residual data; store the processed residual data; add the prediction block to the residual block to determine a reconstructed block; and perform adaptive loop filtering (ALF) on the reconstructed block, based on the processed residual data.