Significance Map Partitioning for Context-Adaptive Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video encoding standards, such as H.264/AVC and developing HEVC, face inefficiencies in encoding and decoding significance maps due to uniform context distribution, which leads to suboptimal compression efficiency and adaptivity, especially as the number of contexts increases with larger transform units.
Innovation Solution
Implementing a non-spatially uniform partitioning of significance maps into parts, where each bit position is associated with a specific context, allowing for context-adaptive encoding and decoding by dynamically selecting and updating contexts based on reconstructed bit values, thereby optimizing compression efficiency and computational speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If uniform context distribution is used in significance map encoding, then the encoding process is simple, but compression efficiency is suboptimal
Solution Approach 1:
The significance map is divided into multiple partitions, with each partition associated with a separate context. This segmentation allows different context models to be applied to different regions, improving compression efficiency by capturing local statistical variations while maintaining manageable complexity through systematic organization.
Solution Approach 2:
Different context models are assigned to different partitions of the significance map, allowing each region to be encoded with a context that best matches its local characteristics. This local adaptation improves compression efficiency by better modeling the statistical properties of transform coefficients in different areas.
2Adaptability or versatility
If more contexts are tracked for better adaptivity, then probability estimates improve, but computational burden increases
Solution Approach 1:
The significance map is divided into a limited number of partitions (e.g., 4 partitions for 4x4 blocks, 2 partitions for 8x8 blocks), which balances adaptivity with computational feasibility. This segmentation provides sufficient local adaptation while keeping the number of contexts manageable for high-speed encoding and decoding.
Solution Approach 2:
Instead of tracking a context for every single coefficient position, the patent applies partial action by grouping coefficients into partitions and using a smaller number of contexts. This reduces computational burden while still capturing the essential local variations needed for improved probability estimates.
3Measurement precision
If separate contexts are associated with each coefficient position, then encoding precision improves, but the number of contexts to track becomes large
Solution Approach 1:
The significance map is segmented into partitions that group multiple coefficient positions together. Each partition is associated with a separate context, providing localized precision without requiring a distinct context for every individual coefficient position. This reduces the total number of contexts while maintaining encoding precision through local adaptation.
Solution Approach 2:
Multiple coefficient positions are merged into partitions that share common contexts. This combining approach reduces the number of contexts to track while still preserving the ability to model local variations, as coefficients within the same partition exhibit similar statistical properties.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
Methods of encoding and decoding for video data are describe in which significance maps are encoded and decoded using non-spatially-uniform partitioning of the map into parts, wherein the bit positions within each part are associated with a given context. Example partition sets and processes for selecting from amongst predetermined partition sets and communicating the selection to the decoder are described.