Video Coding Subblock Merge Mode With DMVR Refinement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies face challenges in efficiently compressing video data, particularly in reducing redundancy and managing motion vectors, which affects bandwidth and storage requirements, especially in applications like video streaming and conferencing.
Innovation Solution
The implementation of decoder-side motion vector refinement (DMVR) and subblock-based merge modes, such as subblock-based temporal motion vector prediction (SbTMVP), to refine motion vectors and improve compression efficiency by analyzing and adjusting motion information at a subblock level within the video encoding and decoding process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If motion compensation with motion vectors is used for video compression, then video quality is maintained, but bandwidth and storage requirements increase significantly
Solution Approach 1:
The current block is divided into multiple subblocks, and each subblock is independently processed with its own motion vector refinement. This segmentation allows for more precise motion compensation at a finer granularity, improving video quality while maintaining compression efficiency by only encoding the differences at subblock level rather than for the entire block.
Solution Approach 2:
Different refinement processes are applied to different subblocks based on their specific characteristics. The decoder-side motion vector refinement (DMVR) and subblock-based temporal motion vector prediction (SbTMVP) allow each subblock to have locally optimized motion compensation parameters, improving overall video quality without uniformly increasing the data requirement for the entire block.
2Productivity
If motion vectors are refined at a finer granularity (subblock level), then compression efficiency improves, but processing complexity increases
Solution Approach 1:
The decoder performs motion vector refinement on its own side without requiring additional information from the encoder. The DMVR technique allows the decoder to independently refine motion vectors using locally available reference pictures and block data, improving compression efficiency while avoiding the complexity of coordinating refinement between encoder and decoder.
Solution Approach 2:
Temporal motion vector prediction is performed in advance by predicting motion vectors for subblocks based on previously decoded temporal information. This preliminary action provides refined motion vectors before the actual decoding process, improving compression efficiency while distributing the processing workload in a manageable sequence.
3Measurement precision
If subblock-based merge modes are used, then motion information accuracy improves, but data required for encoding increases
Solution Approach 1:
Multiple motion vector prediction techniques (DMVR and SbTMVP) are merged into a unified subblock-based merge mode. This combination allows the system to leverage the strengths of both techniques - the local refinement capability of DMVR and the temporal prediction efficiency of SbTMVP - achieving high motion information accuracy while sharing common data structures and processing logic to minimize encoding overhead.
Data Source
AI summary
Aspects of the disclosure provide a method and an apparatus for video coding. The apparatus includes processing circuitry. The processing circuitry can determine a subblock-based merge mode used to code a current block in a current picture. The processing circuitry can partition the current block into a plurality of subblocks based on the subblock-based merge mode. The processing circuitry can determine that decoder-side motion vector refinement (DMVR) is to be applied to one of the plurality of subblocks in the current block based on a mode type of the subblock-based merge mode. The processing circuitry can encode information indicating the subblock-based merge mode in a bitstream.