Multi-Level Significance Maps for Large-Block Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding standards, such as H.264/AVC and HEVC, face inefficiencies in encoding and decoding large transform units like 16×16 and 32×32 due to costly memory access operations and context determination for significance maps, which consume a significant portion of the encoded data.
Innovation Solution
Implement multi-level significance maps by grouping significant-coefficient flags into larger transform units (e.g., 16×16 and 32×32) and using higher-level significance maps (L1 maps) to reduce the number of contexts required, thereby optimizing entropy encoding and decoding processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of substance
If block-based coding with discrete cosine transform and quantization is used, then data compression is achieved, but the encoding and decoding of large transform units becomes computationally expensive due to costly memory access operations and context determination
Solution Approach 1:
The patent divides large transform units (16×16 and 32×32) into smaller 4×4 sub-blocks for processing. Each sub-block has its own significance map and context model, allowing independent encoding/decoding operations. This segmentation reduces the computational burden on large blocks while maintaining compression efficiency through localized processing.
Solution Approach 2:
The patent introduces a hierarchical structure with multiple levels of significance maps. Level-0 significance maps operate at the 4×4 sub-block level, while level-1 significance maps operate at the 16×16 or 32×32 transform unit level. This multi-dimensional approach allows the system to manage complexity across different scales simultaneously.
2Measurement precision
If separate context models are used for each coefficient position in large transform units, then encoding precision is improved, but the number of contexts to track increases significantly
Solution Approach 1:
The patent creates separate context models for each 4×4 sub-block within larger transform units. Instead of tracking contexts for all 256 coefficient positions in a 16×16 block simultaneously, the system segments them into 16 separate 4×4 groups, each with its own context model. This reduces the complexity of context management while preserving encoding precision within each subgroup.
Solution Approach 2:
The patent introduces a hierarchical context model with two levels: level-0 contexts for 4×4 sub-blocks and level-1 contexts for 16×16 or 32×32 transform units. This multi-level structure allows the system to maintain detailed context information where needed while managing overall complexity through hierarchical organization.
3Loss of substance
If significance maps are encoded for large transform units, then compression efficiency is improved, but memory access costs and processing time increase
Solution Approach 1:
The patent divides the significance map encoding task into smaller 4×4 sub-block units. Each sub-block's significance map can be processed independently and in parallel, reducing the sequential processing time required for large transform units while maintaining the compression benefits of significance map encoding.
Solution Approach 2:
The patent performs preliminary organization of coefficient groups into 4×4 sub-blocks before the actual encoding process. This preliminary segmentation allows the encoding algorithm to work with pre-organized data structures, reducing the computational overhead during the actual encoding phase and improving overall processing efficiency.
Data Source
AI summary
Methods of encoding and decoding for video data are described in which multi-level significance maps are used in the encoding and decoding processes. The significant-coefficient flags that form the significance map are grouped into contiguous groups, and a significant-coefficient-group flag signifies for each group whether that group contains no non-zero significant-coefficient flags. If there are no non-zero significant-coefficient flags in the group, then the significant-coefficient-group flag is set to zero. The set of significant-coefficient-group flags is encoded in the bitstream. Any significant-coefficient flags that fall within a group that has a significant-coefficient-group flag that is non-zero are encoded in the bitstream, whereas significant-coefficient flags that fall within a group that has a significant-coefficient-group flag that is zero are not encoded in the bitstream.