Motion Candidate Reordering for Higher-Efficiency Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies, such as MPEG-2, MPEG-4, AVC, HEVC, and VVC, require improvements in motion candidate list construction to enhance coding efficiency.
Innovation Solution
Implement a reordering process for motion candidate lists, including averaging predefined pairs of reordered motion candidates and determining a set of pairwise average motion candidates during video conversions, and utilizing various types of motion candidate lists such as spatial, temporal, history-based, and subblock-based candidates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional motion candidate list construction is used, then the implementation is simple, but coding efficiency is insufficient
Solution Approach 1:
The patent applies a reordering process to the motion candidate list before using it for prediction. By preliminarily organizing the motion candidates in a specific order (prioritizing spatial candidates, then temporal candidates, then HMVP candidates), the system prepares the data structure in advance to enable more efficient selection of optimal motion candidates during encoding, thereby improving coding efficiency without adding complex real-time processing requirements
Solution Approach 2:
The patent segments the motion candidate list into distinct categories: spatial motion candidates, temporal motion candidates, and HMVP candidates. This segmentation allows the system to process and evaluate different types of motion candidates separately, applying appropriate selection criteria to each segment, which improves the overall effectiveness of motion compensation while maintaining a structured and manageable approach
2Productivity
If motion candidate list construction is improved through reordering and pairwise averaging, then coding efficiency increases, but computational complexity increases
Solution Approach 1:
The patent implements pairwise averaging selectively for specific pairs of motion candidates in the reordered list rather than averaging all possible pairs. By applying the averaging operation only to predetermined pairs (e.g., adjacent candidates or candidates with similar characteristics), the system achieves improved prediction accuracy through partial averaging, reducing unnecessary computational overhead while still benefiting from the smoothing effect of averaging in critical cases
Solution Approach 2:
The patent changes the ordering parameters of the motion candidate list by applying a specific reordering scheme that prioritizes candidates based on their source (spatial, temporal, HMVP) and their predicted effectiveness. This parameter change in the organization structure enables more efficient traversal and selection, reducing the computational effort needed to identify optimal motion candidates while improving 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 reordering process to a motion candidate list for the target block; determining a set of pairwise average motion candidates by averaging a set of predefined pairs of reordered motion candidates; and performing the conversion based on the set of pairwise average motion candidates.