Video Decoder Motion Vector Refinement via Pixel Downsampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The high calculation complexity of the decoder-side motion vector refinement (DMVR) technology in video encoding and decoding frameworks hinders efficient picture coding and decoding processes.
Innovation Solution
A method and apparatus that reduce the calculation amount of difference comparison by determining motion information, performing motion search with first precision, and downsampling pixel arrays to identify a target prediction motion vector, thereby improving picture coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If decoder-side motion vector refinement (DMVR) technology is used to remove redundant information, then picture coding efficiency is improved, but calculation complexity increases
Solution Approach 1:
The patent segments the picture block into multiple sub-blocks and performs motion search separately for each sub-block. This segmentation reduces the overall calculation complexity by dividing the large-scale motion search into smaller, more manageable units while maintaining coding efficiency through localized motion compensation.
Solution Approach 2:
The patent applies different motion search strategies to different regions of the picture block. By performing motion search only in necessary sub-blocks and using different search ranges based on local characteristics, the method reduces unnecessary calculations while maintaining high prediction accuracy in critical areas.
2Measurement precision
If motion search with high precision is performed to obtain accurate prediction blocks, then picture quality is improved, but calculation amount increases
Solution Approach 1:
The patent divides the motion search process into multiple precision levels across different sub-blocks. High-precision motion search is performed only where necessary, while other areas use lower precision, thereby reducing the total calculation amount while maintaining overall prediction accuracy.
Solution Approach 2:
The patent performs motion search with full precision only in sub-blocks that require it, rather than uniformly applying high precision across the entire picture block. This partial application of high-precision action reduces unnecessary calculations while maintaining quality where needed.
Data Source
AI summary
A picture decoding method includes determining motion information of a to-be-decoded block; obtaining a first decoding prediction block of the to-be-decoded block based on the motion information; performing motion search with first precision in the prediction reference picture block to obtain at least two second decoding prediction blocks; performing downsampling on the first decoding prediction block and the at least two second decoding prediction blocks to obtain a first sampling pixel array and at least two second sampling pixel arrays, respectively; calculating a difference between the first sampling pixel array and each of the second sampling pixel arrays, and using, as a target prediction motion vector, a motion vector corresponding to a second sampling pixel array with a minimum difference; and obtaining a target decoding prediction block of the to-be-decoded block based on the target prediction motion vector, and decoding the to-be-decoded block based on the target decoding prediction block.