Block-Based Context Selection for Video Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing context selection methods for significance map and significance group flag coding in video coding, such as in HEVC, require complex calculations and operations, particularly for TU sizes other than 4x4 and 8x8, which can lead to inefficiencies in the encoding process due to the need to determine regions and context sets based on coefficient positions.
Innovation Solution
The proposed solution introduces a block-based context selection method that simplifies the process by using sub-block indices in scan order, horizontal and vertical sub-block indices, and video component types to determine context sets, eliminating the need for calculating coefficient positions and unifying context selection for significance map and significance group flag coding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If existing context selection methods are used for significance map and significance group flag coding, then coding accuracy can be maintained, but computational complexity increases and encoding efficiency decreases
Solution Approach 1:
The patent merges the context selection processes for significance map and significance group flag coding into a unified method. Both syntax elements now use the same context set selection based on sub-block indices and scan order, eliminating the need for separate region determination and dual context set management, thereby reducing computational complexity while maintaining coding accuracy
Solution Approach 2:
The patent changes the parameters used for context selection from coefficient positions and region-based indexing to sub-block indices in scan order. This parameter transformation simplifies the context selection process by using straightforward index-based lookup tables instead of complex position calculations and region determinations
2Speed
If complex region determination and dual context sets are used for TU sizes other than 4x4 and 8x8, then coding precision can be maintained, but encoding speed decreases
Solution Approach 1:
The patent segments the transform unit into sub-blocks and uses the sub-block indices to determine context sets. This segmentation approach allows the use of simplified index-based context selection that works uniformly across all TU sizes, eliminating the need for complex region determination while maintaining coding precision through appropriate context model selection
Solution Approach 2:
The patent creates a universal context selection method that works for all TU sizes (4x4, 8x8, 16x16, 32x32, and non-square TUs) using the same sub-block index-based approach. This universal method eliminates the need for size-specific region determination logic while maintaining coding precision through unified context set selection
Data Source
Figure 1~2
Figure 3~4
Figure 5A~5B
AI summary
A method and apparatus for significance group flag coding are disclosed. According to the present invention, the TUs are divided into sub-blocks and significant group flags associated with a TU are coded based on context set selection. The context set selection depends on horizontal sub-block index, vertical sub-block index, or both the horizontal sub-block index and the vertical sub-block index. In one embodiment, two sub-blocks use same context set selection for significance group flag coding if the two sub-blocks use same context set selection for significance map coding.