Intra-Prediction Mode Based Transform Coefficient Transposition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards like AVS face inefficiencies in extending block sizes and fail to utilize side information for intra prediction in residual blocks, leading to suboptimal compression performance.
Innovation Solution
The implementation of block transposition and improved context design for entropy coding, where transform coefficient blocks are treated as if they belong to different prediction classes based on intra-prediction modes, reducing the number of contexts and enhancing coding efficiency by modifying the position encoding of the last non-zero coefficient.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If straightforward extension of the existing AVS coding scheme is applied to larger block sizes, then the coding process remains simple, but coding efficiency deteriorates
Solution Approach 1:
The patent divides the transform coefficient block into multiple sub-blocks and processes each sub-block separately with its own context model. This segmentation allows the coding scheme to adapt to larger block sizes while maintaining efficiency, as each sub-block can be coded independently with optimized context designs.
Solution Approach 2:
The patent introduces dynamic context modeling where context models are selected and updated based on local characteristics of different regions within the block. This dynamic adaptation enables the coding scheme to efficiently handle varying patterns in larger blocks rather than using a static, one-size-fits-all approach.
2Device complexity
If side information like intra prediction information is not used in coding the residual blocks, then the coding process remains simple, but compression performance deteriorates
Solution Approach 1:
The patent utilizes intra prediction mode information as feedback to guide the entropy coding process. The context models are selected and configured based on the predicted intra mode, creating a feedback loop where prediction information directly influences the coding strategy, thereby improving compression without excessive complexity.
Solution Approach 2:
The patent changes coding parameters (context model selection, binarization methods) based on the intra prediction mode. Different prediction modes trigger different coding parameter configurations, allowing the system to adapt to the statistical characteristics of residual blocks generated by each prediction type.
3Productivity
If the number of contexts in entropy coding is increased for better distribution-based design, then compression efficiency improves, but coding complexity increases
Solution Approach 1:
The patent segments the entropy coding process into multiple context models, each responsible for specific regions or types of coefficients. This segmentation allows for better distribution-based design with multiple specialized contexts while managing complexity through modular organization and selective application of different context models.
Data Source
Figure 1a
Figure 1b
Figure 2a
AI summary
Methods and devices for video coding are described. If an intra-prediction mode to be used in encoding a block of residual data is in a specific predetermined class, e.g., the horizontal class, then at least a portion of a quantized transformed block of residual data is transposed during encoding. Likewise, if an intra-prediction mode that was used to generate an encoded block of residual data is in such a class, then at least a portion of an entropy-decoded block of residual data is transposed during decoding.