Geometric Partition Motion Refinement for Lower Signaling Overhead
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards like VVC and AVS3 face inefficiencies in the geometric partition mode (GPM) due to suboptimal motion vector refinement, particularly in non-merge inter modes, leading to increased signaling overhead and reduced coding efficiency.
Innovation Solution
Implementing geometric partition mode with motion vector refinement (GPM-MVR) that applies predefined motion vector refinements on top of existing uni-directional MVs for GPM partitions, using syntax elements similar to MMVD design, and extending GPM mode to explicit inter modes with improved motion signaling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If motion vector refinement is applied to GPM partitions, then motion vector accuracy is improved, but signaling overhead increases
Solution Approach 1:
The patent applies parameter changes by introducing motion vector refinement offsets that modify the base motion vectors. Multiple candidate offset sets are defined with different parameter configurations, allowing the system to change motion vector parameters adaptively to improve accuracy while managing signaling overhead through selective application.
Solution Approach 2:
The patent implements dynamics by enabling adaptive switching between different motion vector refinement offset sets based on content characteristics. The control variable mechanism allows the system to dynamically select appropriate refinement levels, applying motion vector refinement only when beneficial, thus balancing accuracy improvement against signaling overhead.
2Productivity
If adaptive switching between MVR offset sets is implemented, then coding efficiency is improved, but device complexity increases
Solution Approach 1:
The patent applies segmentation by dividing motion vector refinement into multiple discrete offset sets, each optimized for different content scenarios. The control variable segments the decision space, allowing the decoder to select from predefined options rather than computing refinements continuously, thus improving coding efficiency while limiting complexity growth.
Solution Approach 2:
The patent uses cheap short-living objects by employing lightweight control variables that enable rapid switching between offset sets. These control variables require minimal signaling and processing, allowing frequent adaptive decisions without significant complexity overhead, thus improving coding efficiency with acceptable device complexity.
Data Source
AI summary
Methods and devices for video decoding are provided. The method includes: receiving a control variable associated with the video block at a coding level, where the control variable enables adaptive switch between a plurality of motion vector refinement (MVR) offset sets, and the video block comprises first and second geometric partitions; receiving one or more syntax elements to determine a first MVR offset for the first geometric partition and a second MVR offset for the second geometric partition from a selected MVR offset set; obtaining a first motion vector (MV) and a second MV from a candidate list for the first and second geometric partitions; calculating a first refined MV and a second refined MV based on the first and second MVs and the first and second MVR offsets; and obtaining prediction samples for the video block based on the first and second refined MVs.