Bi-Motion Vector Refinement for Video Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies, such as HEVC and VVC, face limitations in achieving optimal coding efficiency due to challenges in accurately refining motion vectors and predicting motion information, which affects the compression and decoding of video data.
Innovation Solution
The implementation of decoder side motion vector refinement (DMVR) processes, including multiple stages of motion vector refinement and adaptive methods, to enhance the prediction and refinement of motion vectors during the video encoding and decoding process, particularly in advanced motion vector prediction modes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional motion vector prediction methods are used, then device complexity is reduced, but measurement precision of motion vectors deteriorates
Solution Approach 1:
The motion vector refinement process is divided into multiple stages: initial motion vector prediction from neighboring blocks, followed by separate refinement of L0 and L1 motion vectors using bilateral matching. This segmentation allows precise motion vector acquisition without requiring a completely complex alternative system.
Solution Approach 2:
The patent performs preliminary motion vector prediction using AMVP mode before the actual bilateral matching refinement. This preliminary action provides a good initial estimate that reduces the search space and computational burden of the subsequent refinement stage, achieving precision without excessive complexity.
2Loss of information
If multiple stages of motion vector refinement are applied, then coding gain is improved, but loss of time increases
Solution Approach 1:
The patent implements a multi-stage refinement process where motion vectors are refined in periodic stages: initial prediction, first refinement stage, and optional second refinement stage. This periodic action allows the system to achieve high coding gain while providing opportunities for early termination if coding targets are met, managing processing time effectively.
Solution Approach 2:
The patent applies motion vector refinement selectively - performing full two-stage refinement only when beneficial, and allowing early termination when coding targets are achieved. This partial action approach avoids unnecessary processing time while maintaining coding efficiency when needed.
3Manufacturing precision
If adaptive DMVR is used, then manufacturing precision of video coding is improved, but device complexity increases
Solution Approach 1:
The patent implements adaptive DMVR that dynamically adjusts the refinement process based on coding conditions. The system can adaptively select between different refinement stages and termination conditions, allowing high precision when needed while reducing complexity for simpler cases through conditional logic.
Solution Approach 2:
The patent uses feedback mechanisms where coding results from intermediate stages are evaluated to determine whether additional refinement is necessary. This feedback-driven approach allows the system to achieve high precision when beneficial while avoiding unnecessary processing complexity when coding targets are already met.
Data Source
AI summary
Embodiments of the present disclosure provide a solution for video processing. A method for video processing is proposed. The method comprises: refining, during a conversion between a target block of a video and a bitstream of the target block, a bi-motion vector predictor (MVP) of the target block by applying a decoder side motion vector derivation process, and wherein the target block is coded with an advanced motion vector prediction (AMVP) mode; and performing the conversion based on the refined bi-MVP.