Mode-Aware HMVP Table Pruning for Video Coding

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing video coding standards face challenges in efficiently managing motion vector prediction and candidate lists, leading to increased computational complexity and reduced coding efficiency.

Innovation Solution

Implementing pruning processes for motion vector prediction and candidate lists based on video characteristics, including enabling or disabling table updates and applying rules for categorizing and pruning motion candidates, to optimize the conversion between video blocks and bitstream representations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If pruning processes are applied to motion vector prediction and candidate lists, then coding efficiency is improved, but computational complexity increases

Engineering Contradiction:

Improvecoding efficiencyVSAvoidcomputational complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent segments the motion vector prediction process into different phases: candidate generation, pruning, and final selection. By dividing the complex task into manageable segments with specific pruning rules for different candidate types (spatial, temporal, HMVP), the system achieves efficient coding while controlling computational complexity through structured processing.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies different pruning strategies to different types of motion candidates based on their local characteristics. Spatial merge candidates use one set of pruning rules, temporal candidates use another, and HMVP candidates use yet another. This localized approach allows optimization of coding efficiency for each candidate type while managing overall computational complexity through targeted processing.

Inventive Principle:

Principle #3Local quality

2Manufacturing precision

If motion candidate lists are extended to improve prediction accuracy, then video quality improves, but processing time increases

Engineering Contradiction:

Improveprediction accuracyVSAvoidprocessing time

Core Design Contradiction:

Manufacturing precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary actions by pre-defining pruning rules and criteria for motion candidates before the actual prediction process. The pruning strategies are established in advance based on candidate types and their characteristics, allowing the system to quickly filter and select optimal candidates during video processing without time-consuming computations.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent applies partial pruning actions rather than complete exhaustive search. By selectively applying pruning rules to specific candidate types and using heuristic-based filtering, the system achieves sufficient prediction accuracy without the computational burden of evaluating all possible candidates, thus reducing processing time while maintaining quality.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS12445602B2Pruning method in different prediction mode

Publication Date: 2025.10.14 DOUYIN VISION CO LTD

AI summary

A method of video processing includes determining, based on a video characteristic of a first video block, an enabling or disabling of a pruning process for a table updating process; updating one or multiple tables, based on the determination and motion information of the first video block, the one or multiple motion tables being History-based Motion Vector Prediction (HMVP) tables; and performing a conversion between a subsequent video block of the video and bitstream representation of the subsequent video block based on the updated tables.