Adaptive Motion Candidate Reordering for Geometric Partitioning Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies, such as HEVC and VVC, face limitations in coding efficiency due to conventional motion candidate list generation processes, which do not adapt effectively to geometric partitioning modes, leading to suboptimal performance in video compression.
Innovation Solution
An adaptive motion candidate list generation process is introduced, specifically for geometric partitioning modes, where motion candidates are reordered to improve the effectiveness of merge mode operations during video processing, enhancing coding efficiency by optimizing motion prediction and compensation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional motion candidate list generation process is used, then device complexity is low, but coding efficiency is limited
Solution Approach 1:
The motion candidate list generation process is made dynamic by adapting the list composition based on the geometric partitioning mode. Different candidate lists are generated for different GPM configurations (e.g., different numbers of partitions), allowing the system to optimize motion prediction for each specific partitioning scenario rather than using a fixed conventional list.
Solution Approach 2:
The patent applies local quality by tailoring the motion candidate list specifically for geometric partitioning modes. Instead of using a universal candidate list for all coding modes, the system generates specialized candidate lists that are locally optimized for GPM operations, improving prediction accuracy for each local coding scenario.
2Reliability
If adaptive motion candidate list generation for GPM is implemented, then motion prediction effectiveness is improved, but processing complexity increases
Solution Approach 1:
The system performs preliminary action by pre-defining the structure and composition of motion candidate lists for different geometric partitioning modes. The candidate list generation rules are established in advance based on the partitioning configuration, allowing efficient runtime selection without complex real-time computation.
Solution Approach 2:
The patent utilizes parameter changes by adjusting the motion candidate list composition based on geometric partitioning parameters (such as the number of partitions). By changing the list parameters adaptively according to the GPM configuration, the system optimizes motion prediction while controlling processing complexity through parameter-driven adaptation rather than structural complexity.
3Productivity
If conventional merge mode is used without reordering, then implementation is simple, but compression performance is suboptimal
Solution Approach 1:
The merge mode operation is made dynamic by reordering motion candidates based on the geometric partitioning mode. The system dynamically adjusts the candidate list order according to the specific GPM configuration, enabling optimized motion prediction for each partitioning scenario while maintaining a relatively simple implementation through rule-based reordering.
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: determining, during a conversion between a target block of a video and a bitstream of the target block, a coding mode applied to the target block; in response to the coding mode being a geometric partitioning mode (GPM), reordering a plurality of motion candidates; and performing the conversion using the reordered plurality of motion candidates.