Adaptive Loop Filtering Across Virtual Boundaries in Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards face challenges in efficiently handling virtual boundaries of coding tree blocks during adaptive loop filtering, leading to suboptimal compression efficiency and complex implementations.
Innovation Solution
The proposed methods involve determining sample classifications across virtual boundaries for adaptive loop filtering, applying filters based on logical groupings of coding tree blocks, and using boundary rules to manage sample usage, thereby simplifying the filtering process and improving compression efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If complex boundary handling and padding methods are used for virtual boundaries during in-loop filtering, then filtering accuracy is improved, but computational requirements increase and pipeline bubbles occur
Solution Approach 1:
The patent segments the boundary handling process by introducing virtual boundaries that divide the picture into regions. These virtual boundaries allow the filtering process to handle different areas with different rules, enabling more accurate filtering at boundaries while avoiding the need for complex padding operations across the entire picture. The segmentation is achieved through the virtual boundary mechanism that creates distinct processing zones without requiring additional computational overhead.
Solution Approach 2:
The patent uses virtual samples as an intermediary mechanism to handle boundary conditions. Instead of directly dealing with complex boundary cases that require padding and multiple processing steps, the virtual samples act as a mediator that simplifies the boundary handling process. These virtual samples are generated based on simple rules and allow the filtering process to proceed without pipeline bubbles or complex computational operations.
2Measurement precision
If multiple padding methods are used for different boundary types, then boundary handling accuracy is improved, but device complexity increases
Solution Approach 1:
The patent applies a unified padding method that works for all boundary types (picture boundaries, slice boundaries, tile boundaries, and virtual boundaries). This universal approach eliminates the need for multiple different padding methods for different boundary types, thereby reducing device complexity while maintaining boundary handling accuracy. The same padding mechanism is applied consistently across all boundary scenarios, simplifying the overall filtering process.
3Productivity
If virtual samples are used for in-loop filtering, then compression efficiency is improved, but computational overhead increases
Solution Approach 1:
The patent applies virtual samples only partially - specifically only at boundary locations where they are needed for accurate filtering, rather than applying them throughout the entire picture. This partial application of virtual samples maintains compression efficiency improvements at critical boundary areas while minimizing the overall computational overhead associated with virtual sample generation and processing.
Data Source
AI summary
A method of video processing includes determining a classification of samples of a block in a video picture of a video according to a rule, wherein the video picture is divided into multiple regions, and wherein the rule disallows use of neighboring samples of a current sample across one or more boundaries of the multiple regions for the classification. The one or more boundaries of the multiple regions comprises a virtual boundary for an adaptive loop filtering (ALF) process. The method also includes performing a conversion between the block and a bitstream representation of the video by selectively applying the ALF process according to the classification.