Video Block Coding With Decoder-Side Motion Vector Refinement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies, such as MPEG-2, MPEG-4, AVC, HEVC, and VVC, require improvements in coding efficiency to enhance digital video capabilities.
Innovation Solution
Implementing decoder side motion vector refinement (DMVR) and its variants during the conversion between a target block of a video and its bitstream, allowing for improved coding efficiency and higher coding gain.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional video coding technologies (HEVC, VVC) are used, then the basic video compression function is achieved, but coding efficiency is limited due to inaccurate motion vector refinement
Solution Approach 1:
The patent applies preliminary action by performing motion vector refinement at the decoder side using reference pictures that have been pre-decoded and stored in a reference picture buffer. The refined motion vectors are calculated before final prediction, allowing the system to leverage previously processed information to improve current coding efficiency without requiring additional encoder complexity.
Solution Approach 2:
The patent implements feedback mechanisms through decoder side motion vector refinement (DMVR) where motion vectors are iteratively adjusted based on decoded reference pictures. The refined motion vectors are then fed back into the prediction process, creating a feedback loop that continuously improves prediction accuracy and coding efficiency without requiring encoder-side changes.
2Productivity
If decoder side motion vector refinement (DMVR) is applied to all blocks, then coding efficiency is maximized, but processing complexity and computational load increase
Solution Approach 1:
The patent applies local quality by selectively applying DMVR to specific blocks or regions within a picture based on their importance, motion characteristics, or coding mode. Not all blocks require the same level of refinement, so the system adapts the refinement process to local requirements, improving coding efficiency where needed while reducing complexity where simple prediction suffices.
Solution Approach 2:
The patent implements dynamics by making the DMVR process adaptive and configurable rather than fixed. The system can dynamically adjust the refinement process based on picture characteristics, block types, and coding conditions. This allows the complexity to be dynamically scaled up or down depending on the specific processing requirements of each frame or block.
3Measurement precision
If motion vector refinement is performed with high precision, then prediction accuracy improves, but processing time and computational resources increase
Solution Approach 1:
The patent applies partial action by performing motion vector refinement to the extent necessary for achieving adequate prediction accuracy, rather than always applying maximum refinement. The system evaluates whether full DMVR is needed for each block and can apply reduced refinement or skip refinement when the benefit is minimal, thus avoiding excessive computational time while maintaining sufficient precision for coding efficiency.
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: applying, during a conversion between a target block of a video and a bitstream of the target block, a process of at least one of: a decoder side motion vector refinement (DMVR) or a DMVR variant to the target block that is coded with a target mode; and performing the conversion based on the process of at least one of: the DMVR or the DMVR variant.