Video Signal Encoding Sub-Region Position Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression techniques face challenges in efficiently coding the position of the last non-zero transform coefficient, particularly in transform units that are divided into sub-regions, which affects the encoding efficiency and data compression.
Innovation Solution
A method is proposed that involves dividing the current block into sub-regions and determining the last non-zero region based on a syntax element. This method adaptively changes the coding method depending on various conditions and considers the presence of the last non-zero coefficient within a limited region, effectively encoding the positional information of the last non-zero coefficient.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If the position of the last non-zero transform coefficient is coded using conventional entropy coding methods, then the coding can be performed sequentially, but the encoding complexity increases and parallel operation becomes difficult
Solution Approach 1:
The transform block is divided into multiple sub-regions, and the position coding is segmented into multiple syntax elements (lastPositionInSubregion, subregionIndex, etc.). This segmentation allows different parts of the position information to be coded independently, enabling parallel processing while maintaining sequential compatibility.
Solution Approach 2:
The position coding is transformed from a single-dimensional sequential code into a multi-dimensional structure involving subregion indices, positions within subregions, and significant coefficient flags. This dimensional expansion enables parallel computation paths while preserving the ability to perform sequential coding when needed.
2Measurement precision
If the transform block is divided into sub-regions for position coding, then the precision of position information is improved, but the amount of data to be coded increases
Solution Approach 1:
Different coding strategies are applied to different sub-regions based on their characteristics. The significantCGflag syntax element locally indicates whether each sub-region contains non-zero coefficients, allowing the encoder to adaptively code only the necessary position information for each sub-region, thereby improving precision without proportionally increasing data amount.
Solution Approach 2:
The coding system uses a hierarchical approach where only the necessary level of detail is coded. If a sub-region is determined to contain no significant coefficients (through significantCGflag), detailed position coding for that sub-region is omitted, achieving partial action that balances precision with data efficiency.
3Productivity
If context-based adaptive binary arithmetic coding is used, then the compression efficiency is improved, but the sequential structure prevents parallel operation
Solution Approach 1:
The CABAC process is segmented into multiple independent context models corresponding to different sub-regions and syntax elements. Each context model can be updated and operated independently, enabling parallel execution of coding operations across different segments while maintaining the compression efficiency benefits of context-based adaptive coding.
Solution Approach 2:
Context models are pre-configured for different sub-regions and syntax element types before the actual coding process. This preliminary setup allows the decoder to prepare multiple context models in advance, enabling parallel decoding operations while maintaining compression efficiency through context-based adaptation.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present disclosure provides a method for decoding a video signal, the method comprising the steps of: decoding, from a bitstream, a syntax element indicating the last non-zero region, wherein the last non-zero region represents a region including the last non-zero transform coefficient in a scan order; splitting a current block into multiple sub-regions; and based on the syntax element, determining the last non-zero region of the current block among the split sub-regions.