Multi-HMVP Motion Vector Prediction for Better Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards like VVC face challenges in efficiently utilizing motion vector prediction, leading to suboptimal coding efficiency and quality due to limitations in updating and diversity of motion vector candidates.
Innovation Solution
Implementing multiple history-based MVP (Motion Vector Prediction) tables with diverse updating rules and insertion strategies to enhance the diversity and relevance of motion vector candidates, including spatial and temporal considerations, to improve coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a single history-based MVP table is used, then the structure is simple, but the diversity and relevance of motion vector candidates are limited
Solution Approach 1:
The patent divides the single MVP table into multiple tables (first HMVP table, second HMVP table, etc.), each maintaining separate motion vector candidates. This segmentation allows the system to store diverse motion vectors with different updating rules and insertion strategies, thereby improving candidate diversity without requiring a single complex table structure.
Solution Approach 2:
The patent introduces an additional dimension by creating multiple HMVP tables instead of using a single table. This dimensional expansion allows the system to organize motion vector candidates in multiple separate structures, each with its own updating rules,ไป่ enhancing the diversity and relevance of candidates while maintaining manageable structural complexity.
2Reliability
If motion vector candidates are updated frequently, then the relevance is improved, but the computational complexity increases
Solution Approach 1:
The patent applies different updating rules to different HMVP tables based on local characteristics. Each table can have customized updating strategies (e.g., different insertion orders, different criteria for adding candidates) that are optimized for specific motion patterns or video content types, thereby improving relevance without uniformly increasing complexity across the entire system.
Solution Approach 2:
The patent implements dynamic updating mechanisms where the system can adaptively select which tables to update and how to update them based on current video content characteristics. This dynamic approach allows frequent updates to maintain relevance while managing computational complexity through selective updating rather than uniform updates across all tables.
3Productivity
If multiple HMVP tables are implemented, then coding efficiency is improved, but the processing time increases
Solution Approach 1:
The patent implements partial action by selectively using only the necessary number of HMVP tables and only updating tables when beneficial. The system can configure the number of tables and their updating frequency based on video content characteristics, applying updates partially rather than continuously, thereby improving coding efficiency while minimizing processing time overhead.
Solution Approach 2:
The patent utilizes parameter changes by adjusting the number of HMVP tables, updating frequencies, and insertion strategies based on video content characteristics. This allows the system to optimize the balance between coding efficiency and processing time by adapting parameters dynamically rather than using fixed settings, thereby improving productivity without excessive time loss.
Data Source
AI summary
Methods for video coding using multiple History-based MVP tables. According to one method, motion information for the current block is derived from the coded data at a decoder according to Merge candidate list or an AMVP (MVP (Adaptive Motion Vector Prediction)) list. The multi-HMVP MVP tables are updated according to the motion information derived for the current block. The Merge candidate list or the AMVP list is updated by inserting one or more MVP candidates from the multi-HMVP MVP tables. According to another method, the multiple HMVP lookup tables are updated according to the motion information derived for the current block, and the multiple HMVP lookup tables are updated according to different updating rules.