Geometric Partition Index Unification for Video Block Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards like HEVC and VVC face inefficiencies in geometric partitioning modes, particularly in GPM, IBC-GPM, and SGPM, which are not optimized for compression efficiency and hardware implementation.
Innovation Solution
Harmonizing geometric partitioning modes (GPM, IBC-GPM, and SGPM) by using a unified set of partition indices defined by a single table, allowing more angles and displacements, and incorporating size-dependent partitions to improve compression efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple separate partition index tables are used for different geometric partitioning modes (GPM, IBC-GPM, SGPM), then each mode can be optimized independently, but the overall device complexity and implementation difficulty increase
Solution Approach 1:
The patent merges multiple separate partition index tables for different geometric partitioning modes (GPM, IBC-GPM, SGPM) into a single unified partition index table. This consolidation reduces the complexity of managing multiple tables while maintaining the ability to support all geometric partitioning modes, directly resolving the contradiction between adaptability and device complexity.
Solution Approach 2:
The unified partition index table is designed to serve multiple geometric partitioning modes simultaneously. Each partition index in the table is associated with angle and distance indices that can define partitions for any of the supported modes, making the table universal and multi-functional rather than mode-specific.
2Device complexity
If a unified partition index table is used for all geometric partitioning modes, then device complexity is reduced and implementation is simplified, but the ability to optimize each mode independently is limited
Solution Approach 1:
The unified partition index table incorporates local quality by allowing different partition configurations (different angles and distances) to be selected for different geometric partitioning modes based on their specific requirements. Each mode can utilize the partition indices that best suit its characteristics while sharing the same table structure.
Solution Approach 2:
The system dynamically selects appropriate partition indices from the unified table for each geometric partitioning mode based on the current coding context. This dynamic selection allows the system to adapt to different mode requirements without needing separate static tables for each mode.
3Productivity
If more angles and displacements are incorporated in geometric partitioning, then compression efficiency is improved, but the complexity of the partitioning system increases
Solution Approach 1:
The partitioning system is segmented into discrete partition indices, each representing a specific angle and distance configuration. This segmentation allows the system to offer multiple partitioning options (improving compression efficiency) while managing complexity through structured indexing rather than continuous parameter management.
Solution Approach 2:
The system manages the increased number of angles and displacements by parameterizing them through angle indices and distance indices. This parameterization allows the incorporation of more partitioning options without proportionally increasing system complexity, as the parameters are managed through compact indexing mechanisms.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A decoding method is disclosed. It is determined that a current block is coded with one geometric partitioning mode of a set of geometric partitioning modes. A partition index is obtained for the current block in a set of partition indices. Each partition index of the set is associated with an angle index and a distance index, the angle index and the distance index defining a partition of the current block in at least two partitions. The current block is then decoded based on the angle index and the distance index associated with the partition index. The set of partition indices is identical for all geometric partitioning modes of the set of geometric partitioning modes.