Loop Filtering Method for Video Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Adaptive loop filtering in video coding involves high computation complexity due to repeated calculations of filtering coefficients and decision updates, which affects compression efficiency and reconstruction quality.
Innovation Solution
A loop filtering method that determines a reference image block's filtering probability and spatial coding information to make decisions based on preset conditions, reducing unnecessary calculations by using pre-trained machine learning models and spatial coding characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If adaptive loop filtering is performed with multiple iterations of covariance calculation and filtering coefficient determination, then reconstruction quality is improved, but computation complexity increases
Solution Approach 1:
The patent performs preliminary classification of image blocks into categories (e.g., flat, edge, texture) before applying filtering coefficients. This preliminary action avoids unnecessary complex calculations for blocks that don't require intensive filtering, while still maintaining high reconstruction quality for blocks that do need it.
Solution Approach 2:
The patent applies different filtering strategies and coefficients to different types of image blocks based on their local characteristics. Simple blocks use minimal filtering while complex blocks receive more intensive processing, optimizing the balance between quality and computation for each local region.
2Measurement precision
If filtering coefficient group is calculated and decision is updated multiple times through Rate-Distortion Optimization, then loop filtering accuracy is improved, but processing time increases
Solution Approach 1:
The patent performs preliminary classification of image blocks into categories before applying filtering coefficients. This preliminary action avoids unnecessary complex calculations for blocks that don't require intensive filtering, while still maintaining high reconstruction quality for blocks that do need it.
Solution Approach 2:
The patent changes the parameter of filtering coefficient selection based on image block characteristics. Different coefficient groups are selected for different block types, allowing accurate filtering to be applied only where needed rather than uniformly across all blocks.
3Manufacturing precision
If loop filtering is applied to all image blocks to reduce compression distortion, then reconstruction quality is improved, but compression efficiency decreases
Solution Approach 1:
The patent applies different filtering strategies and coefficients to different types of image blocks based on their local characteristics. Simple blocks use minimal filtering while complex blocks receive more intensive processing, optimizing the balance between quality and computation for each local region.
Solution Approach 2:
The patent changes the parameter of filtering coefficient selection based on image block characteristics. Different coefficient groups are selected for different block types, allowing accurate filtering to be applied only where needed rather than uniformly across all blocks.
Data Source
AI summary
Embodiments of the present disclosure provide a loop filtering method, including: determining a reference image block corresponding to a to-be-processed image block in an adjacent reference frame in a time domain, and predicting the loop filtering enabling probability of the to-be-processed image block based on the result of loop filtering of the reference image block; determining spatial coding information of the to-be-processed image block, the spatial coding information being used to characterize texture complexity of the to-be-processed image block; and making a loop filtering decision for the to-be-processed image block in response to the spatial coding information and the loop filtering enabling probability meeting preset filtering conditions.