Affine Motion Prediction for Small-Block Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards like HEVC only apply translation motion models, which are inadequate for complex real-world motions such as zoom, rotation, and perspective motions, while VVC introduces block-based affine transform motion compensation but limits it to certain block sizes, preventing efficient application on smaller blocks.
Innovation Solution
Adapting affine motion compensation prediction to allow smaller block sizes by using flexible subblock sizes based on chroma sampling formats and simplifying syntax elements for affine Merge and AMVP modes on smaller coding units, enabling the use of 4-parameter affine models for reduced bit signaling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If block-based affine transform motion compensation prediction is applied to all block sizes, then motion modeling accuracy is improved, but computational complexity and processing overhead increase
Solution Approach 1:
The patent applies different affine MCP mode candidate sets based on local block size characteristics. Large blocks (both dimensions > threshold) receive the first candidate set with more modes, while small blocks (at least one dimension ≤ threshold) receive the second candidate set with fewer modes. This local differentiation optimizes motion modeling accuracy where needed while reducing complexity where sufficient performance is achieved with simpler models.
2Measurement precision
If more affine MCP mode candidates are provided for large blocks, then prediction accuracy is improved, but bit complexity increases
Solution Approach 1:
The patent implements local quality by providing different numbers of affine MCP mode candidates based on block size. The first affine MCP mode candidate set contains more candidates for large blocks where higher prediction accuracy is beneficial, while the second set contains fewer candidates for small blocks where the overhead would outweigh the benefits. This selective approach optimizes the balance between prediction accuracy and bit complexity.
3Productivity
If affine motion compensation is limited to certain block sizes, then processing efficiency is maintained, but applicability to small blocks is reduced
Solution Approach 1:
The patent introduces dynamics by making the affine MCP mode candidate set configurable based on block size characteristics. Rather than a fixed limitation, the system dynamically selects between two candidate sets (first for large blocks, second for small blocks) based on whether block dimensions exceed a threshold. This dynamic adaptation maintains processing efficiency through appropriate candidate set selection while extending versatility to support affine motion compensation on small blocks.
Data Source
AI summary
A method for performing affine motion compensation prediction (MCP) in a video decoder is provided. The method includes selecting an affine MCP mode from an affine MCP mode candidate set. The method also includes performing, based on the selected affine MCP mode, affine motion compensation prediction in the video decoder. For a coding unit with both a width and a height larger than a first threshold, a first affine MCP mode candidate set is applied. For a coding unit with at least one of a width and a height not larger than the first threshold, a second affine MCP mode candidate set is applied. The first affine MCP mode candidate set has more affine MCP mode candidates than the second affine MCP mode candidate set.