Affine Merge Motion Vector Differencing for Sub-Block Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video compression techniques struggle to efficiently manage bandwidth demand for digital video, particularly as the number of connected devices increases, leading to growing bandwidth requirements.
Innovation Solution
The implementation of sub-block based motion vector refinement in video coding and decoding processes, which involves modifying affine motion information in bitstream representations using motion vector differences and/or reference picture indices during decoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional video compression techniques are used, then bandwidth usage is reduced, but video quality and compression efficiency deteriorate as the number of connected devices increases
Solution Approach 1:
The current video block is divided into multiple sub-blocks, and motion vector refinement is performed independently for each sub-block. This segmentation allows for more precise motion compensation at a finer granularity level, improving compression efficiency without requiring excessive bandwidth
Solution Approach 2:
Motion vector refinement is applied selectively to different sub-blocks based on their local characteristics. Each sub-block can have its own refined motion vector, allowing the coding system to adapt to local motion variations and improve overall compression performance
2Productivity
If sub-block based motion vector refinement is implemented, then video coding efficiency is improved, but computational complexity increases
Solution Approach 1:
Affine merge candidates are constructed and prepared in advance before the actual encoding/decoding process. The candidate list is built using motion information from neighboring blocks, so that during actual coding, the decoder can directly select from pre-computed candidates without performing complex calculations
Solution Approach 2:
Motion vectors and affine parameters are copied from neighboring blocks to form merge candidates. Instead of computing motion vectors from scratch for each block, the system reuses motion information from spatially or temporally neighboring blocks, significantly reducing computational complexity
3Measurement precision
If affine motion information is modified using motion vector differences, then motion accuracy is improved, but bitstream complexity increases
Solution Approach 1:
Motion vector differences are applied selectively to refine motion vectors only where needed, rather than applying full precision motion vectors everywhere. This partial refinement approach improves motion accuracy in regions requiring it while keeping the bitstream complexity manageable by not over-refining all blocks
Data Source
AI summary
A method of visual media processing, includes: performing a conversion between a current video block of a visual media data and a bitstream representation of the visual media data according to a modification rule, wherein the current video block is coded using affine motion information; wherein the modification rule specifies to modify an affine motion information coded in the bitstream representation using motion vector differences and/or reference picture indices during decoding.