Regional Context Selection for Transform Unit Significance Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video encoding standards, such as H.264/AVC and developing MPEG-H, face inefficiencies in encoding and decoding significance maps due to the large number of contexts required for larger transform units, leading to high bitstream occupancy and computational complexity.
Innovation Solution
Implementing multi-level significance maps that partition transform units into regions based on coefficient groups, using context-adaptive encoding and decoding to reduce the number of contexts needed, particularly by dynamically determining region boundaries based on neighboring significant-coefficient-group flags.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a separate context is associated with each coefficient position in the transform unit, then encoding accuracy is improved, but device complexity increases due to tracking and looking up a large number of different contexts
Solution Approach 1:
The transform unit is divided into multiple regions (e.g., first region and second region) based on coefficient group flags. Different context sets are assigned to different regions, reducing the total number of contexts from one per coefficient position to a manageable set of regional contexts. This segmentation approach maintains encoding accuracy while significantly reducing context management complexity.
Solution Approach 2:
Different context sets are applied to different regions of the transform unit based on the characteristics of coefficient groups in each region. The DC coefficient position uses a first context set, while other positions use a second context set, allowing locally optimized encoding parameters for different spatial locations within the transform unit.
2Device complexity
If the number of contexts is reduced for larger transform units, then device complexity is decreased, but encoding precision may be compromised
Solution Approach 1:
The transform unit is segmented into multiple regions with different context sets. This allows the system to use a limited number of context sets (first and second context sets) while still providing differentiated encoding parameters for different regions, thereby maintaining encoding precision without requiring a large number of individual contexts for each coefficient position.
Solution Approach 2:
Instead of providing a separate context for every coefficient position (excessive action), the patent applies context sets partially to regional groups of coefficients. This partial application of context management achieves sufficient encoding precision for practical purposes while significantly reducing the overall number of contexts that must be tracked and managed.
3Productivity
If context selection is optimized by partitioning into regions, then computational complexity is reduced, but the complexity of determining region boundaries increases
Solution Approach 1:
The transform unit is pre-partitioned into multiple regions based on coefficient group flags before the actual encoding process. The region boundaries are determined in advance using the significant-coefficient-group flags, which are already available from the significance map. This preliminary partitioning allows for fast context selection during encoding without adding significant computational complexity.
Solution Approach 2:
The region partitioning scheme uses the significant-coefficient-group flags themselves to determine context set selection. The flags automatically provide the information needed to identify which region a coefficient belongs to, eliminating the need for separate, complex boundary detection algorithms. The system uses its own output (the flags) to drive the context selection process.
Data Source
AI summary
Methods of encoding and decoding for video data are described for encoding or decoding multi-level significance maps. Distinct context sets may be used for encoding the significant-coefficient flags in different regions of the transform unit. In a fixed case, the regions are defined by coefficient group borders. In one example, the upper-left coefficient group is a first region and the other coefficient groups are a second region. In a dynamic case, the regions are defined by coefficient group borders, but the encoder and decoder dynamically determine in which region each coefficient group belongs. Coefficient groups may be assigned to one region or another based on, for example, whether their respective significant-coefficient-group flags were inferred or not.