Transform Coefficient Block Coding for Adaptive VLC Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing variable length coding methods for image compression, such as 2D-VLC, are inefficient in encoding sequences of consecutive non-zero-valued coefficients and fail to adapt to the dynamic distribution of coefficients in low-bit rate interframe coding, leading to suboptimal compression and potential litigation issues with existing patents.
Innovation Solution
A method that classifies quantized transform coefficients into multiple classes based on their distribution, applying specific coding processes to each class to optimize bitstream compression, including separate encoding for low-frequency and high-frequency regions, and using run-length and amplitude coding methods tailored to the statistical characteristics of each class.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a single 2D-VLC table is used for encoding quantized transform coefficients, then the coding process is simple and uniform, but the compression efficiency is suboptimal for sequences of consecutive non-zero-valued coefficients and does not adapt to dynamic coefficient distributions
Solution Approach 1:
The patent divides the encoding process into multiple segments by classifying coefficient blocks into different classes (e.g., Class 0 for all-zero blocks, Class 1 for blocks with non-zero coefficients). Each class has its own specialized coding process, allowing the system to handle different coefficient patterns optimally without using a single complex table for all cases.
Solution Approach 2:
The patent introduces dynamic adaptation by selecting different coding processes based on the actual coefficient distribution in each block. The classifier dynamically determines which coding process to apply, enabling the system to adapt to varying coefficient patterns rather than using a static, uniform encoding approach.
2Productivity
If conventional 2D-VLC methods are used, then the encoding process is straightforward, but sequences of consecutive non-zero-valued coefficients are not efficiently encoded
Solution Approach 1:
The patent applies different coding strategies to different local patterns of coefficients. For example, Class 0 blocks (all zeros) use a simple end-of-block code, while Class 1 blocks use run-length encoding for sequences of zeros followed by amplitude coding for non-zero values. This localized optimization improves compression for specific coefficient patterns without complicating the overall process.
3Adaptability or versatility
If a single variable length coding method is applied to all coefficient blocks, then the implementation is uniform and simple, but the method fails to adapt to the dynamic distribution of coefficients in low-bit rate interframe coding
Solution Approach 1:
The patent performs preliminary classification of coefficient blocks before encoding, dividing them into distinct classes based on their characteristics (e.g., all-zero vs. non-zero coefficients). This preliminary action allows the system to prepare and select the appropriate coding process in advance, enabling adaptation to different coefficient distributions without adding complexity during the actual encoding phase.
Data Source
AI summary
A method and apparatus to variable length code an ordered series of quantized transform coefficients of an image block. The method classifies the series into one of a plurality of classes. Each class has a corresponding coding process that is applicable to the statistical distribution of the coefficients in the series. The series is coded according to the particular corresponding coding process of the class of the classifying.