Transform Coefficient Block Coding by Size and Distribution Class
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current digital video compression methods, such as JPEG, MPEG1, and H.264/AVC, use fixed entropy coding for all block sizes within a picture, which can lead to suboptimal compression efficiency due to the uniform application of variable length coding methods regardless of block size and coefficient distribution.
Innovation Solution
The method involves classifying series of quantized transform coefficients by block size and distribution patterns, applying different entropy coding methods to each class, including hybrid and context-adaptive variable length coding techniques, to optimize compression based on the specific characteristics of each block.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a single variable length coding method is applied to all block sizes, then the coding process is simple and uniform, but compression efficiency is suboptimal
Solution Approach 1:
The patent segments the coding process by dividing transform coefficients into different groups based on their distribution characteristics (e.g., first group with more non-zero coefficients, second group with fewer non-zero coefficients). Different variable length coding methods are applied to each group, allowing optimized compression for each segment while maintaining overall system efficiency.
Solution Approach 2:
The patent applies different coding strategies to different regions of the transform coefficient data based on local characteristics. Specifically, context-adaptive binary arithmetic coding (CABAC) is applied to the first group where it provides better compression, while context adaptive variable length coding (CAVLC) is applied to the second group where it is more efficient, matching the coding method to the local data characteristics.
2Loss of energy
If multiple entropy coding methods are applied to different block classes, then compression efficiency is improved, but device complexity increases
Solution Approach 1:
The patent performs preliminary classification of transform coefficients into different groups based on their distribution characteristics before applying the appropriate coding method. This pre-grouping step simplifies the subsequent coding process by ensuring that each group receives the most appropriate coding treatment, reducing the complexity of real-time decision-making during encoding.
Solution Approach 2:
The patent changes the coding parameters (which entropy coding method to use) based on the characteristics of the coefficient groups. By adjusting the coding method parameter according to the distribution patterns of different coefficient groups, the system achieves optimal compression without requiring complex adaptive mechanisms, as the parameter selection is based on pre-determined group characteristics.
Data Source
AI summary
Classifying a series of quantized transform coefficients of a block of image data into one of a pre-defined plurality of classes for entropy coding, and entropy coding the block. A class is defined by at least the size of the block and typically but not necessarily one or more other factors. The classified series is coded by one of a set of pre-defined entropy coding methods, e.g., variable length coding methods for the pre-defined classes.