Multi-pass Decoder Motion Vector Refinement Range Constraint
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies, such as Versatile Video Coding (VVC), face challenges in efficiently refining motion vectors at the decoder side, which can impact prediction accuracy and coding efficiency.
Innovation Solution
The proposed method constrains multi-pass decoder-side motion vector refinement (MP-DMVR) by defining a refinement range for motion vectors, ensuring that only fractional parts of the motion vector are refined while maintaining the integer portion. This refinement is applied in multiple passes, with different constraints for each pass, and is signaled at various levels such as sequence, picture, and slice.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multi-pass decoder-side motion vector refinement is performed without constraints, then prediction accuracy is improved, but computational complexity and memory access increase
Solution Approach 1:
The patent applies parameter changes by constraining the motion vector refinement range to only fractional parts (e.g., ±1/4 pixel) while keeping integer parts fixed. This parameter constraint reduces the search space and computational complexity of multi-pass DMVR while maintaining prediction accuracy improvements.
Solution Approach 2:
The patent segments the motion vector into integer and fractional parts, applying different refinement strategies to each. The integer part remains fixed from the initial motion vector, while only the fractional part undergoes multi-pass refinement. This segmentation reduces computational complexity by eliminating redundant searches.
2Measurement precision
If multi-pass decoder-side motion vector refinement is performed without constraints, then prediction accuracy is improved, but memory access increases
Solution Approach 1:
By constraining the refinement range to a small fractional range (e.g., ±1/4 pixel) around the initial motion vector position, the patent significantly reduces the number of reference picture samples that need to be accessed from memory, while still achieving prediction accuracy improvements through multi-pass refinement.
3Productivity
If refinement range is constrained to maintain integer portion, then coding efficiency is improved, but refinement flexibility is reduced
Solution Approach 1:
The patent changes the refinement parameter scope from the entire motion vector to only its fractional part. This parameter constraint improves coding efficiency by reducing computational complexity while the multi-pass refinement process maintains sufficient flexibility to achieve accurate motion compensation within the constrained range.
Solution Approach 2:
The patent applies partial action by performing refinement only on the fractional part of the motion vector rather than the entire vector. This partial refinement is sufficient to achieve coding efficiency improvements without requiring full motion vector flexibility.
Data Source
AI summary
A method for constraining multi-pass decoder-side motion vector refinement (MP-DMVR) is provided. A video coder receives data for a block of pixels to be encoded or decoded as a current block of a current picture of a video. A video coder receives a motion vector that references a block of pixels in a reference picture based on the received data. A video coder refines the motion vector in a plurality of refinement passes by examining pixels in the reference picture that are identified based on the refined motion vector. The refinement of the motion vector is constrained by a refinement range. The video coder encodes or decodes the current block by using the refined motion vector to produce prediction residuals or to reconstruct the current block.