CTU-Based HMVP Tables for Higher Video Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies, such as VVC, face challenges in efficiently utilizing History-Based Merge Mode and Non-Adjacent Motion Vector Prediction to enhance coding efficiency and reduce bandwidth requirements.
Innovation Solution
Implementing multiple CTU-based History-Based Motion Vector Prediction (MVP) tables with diverse updating rules and constrained FIFO mechanisms to manage and utilize motion information more effectively, including spatial and non-adjacent MVP candidates, to improve coding efficiency and reduce bandwidth.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional single HMVP table is used, then device complexity is low, but coding efficiency is insufficient
Solution Approach 1:
The patent divides the single HMVP table into multiple CTU-based HMVP tables, where each table is associated with a specific CTU (Coding Tree Unit) and stores motion information specific to that region. This segmentation allows the system to maintain lower complexity within each individual table while achieving higher overall coding efficiency through region-specific motion prediction.
Solution Approach 2:
The patent introduces a new dimension of organization by grouping HMVP candidates not just temporally but also spatially according to CTU boundaries. This multi-dimensional organization (temporal + spatial) enables the system to leverage both temporal correlation and spatial locality of motion vectors, significantly improving coding efficiency without proportionally increasing complexity.
2Productivity
If multiple CTU-based HMVP tables are implemented, then coding efficiency is improved, but device complexity increases
Solution Approach 1:
Each CTU-based HMVP table serves multiple functions: it stores motion information for its associated CTU, provides spatial and temporal MVP candidates, and can be selectively accessed based on the current coding context. This multi-functionality justifies the increased number of tables by maximizing their utility in the prediction process.
Solution Approach 2:
The patent applies local quality by creating HMVP tables that are optimized for specific CTU regions. Each table contains motion information tailored to its local area, allowing the system to exploit local motion characteristics more effectively. This local optimization improves coding efficiency while keeping each individual table relatively simple in structure.
3Reliability
If constrained FIFO mechanism is applied to HMVP table updates, then reliability of motion prediction is improved, but processing complexity increases
Solution Approach 1:
The constrained FIFO mechanism implements a feedback loop where motion information is selectively added to and removed from HMVP tables based on predefined constraints. The table update process considers the current CTU position, motion vector characteristics, and historical data to determine when to insert or evict candidates, ensuring reliable prediction while managing complexity through structured feedback control.
Data Source
AI summary
Methods for video coding using CTU-based or multiple History-based MVP (HMVP) tables. According to one method, blocks in the current CTU are encoded or decoded using information comprising a merge list or an AMVP list, where one or more candidates from one or more CTU-based HMVP tables. The CTU-based HMVP tables are maintained and updated on a CTU basis. According to another method, if a to-be-referenced position is not inside the first region, the to-be-referenced position is mapped to a mapped position of the first region before referencing corresponding motion. If the mapped position of the first region has no motion, a predefined default motion or neighbouring motion at a neighbouring position is used as the corresponding motion.