Grouped-Symbol Entropy Coding for Faster Binary Sequence Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current entropy coding methods, such as CABAC in the H.264 standard, are computationally demanding, making them inefficient for devices with limited processing power, particularly in high-quality video decoding applications.
Innovation Solution
The proposed solution involves using encoding trees to generate codewords, where leaves of the trees are associated with codewords from a codeword set, and processing multiple symbols as a group to improve throughput and speed, employing a Huffman code set construction that balances constraints like sequence probability and tree structure efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If CABAC (Context Adaptive Binary Arithmetic Coding) is used for entropy coding, then compression performance is improved, but computational complexity increases significantly
Solution Approach 1:
The patent segments the binary sequence into multiple symbol groups (e.g., groups of 2, 3, or 4 binary symbols) and processes each group as a single unit. This segmentation reduces the number of coding operations required compared to processing each binary symbol individually, thereby lowering computational complexity while maintaining compression effectiveness through context-adaptive coding of the grouped symbols.
Solution Approach 2:
The patent changes the parameter of symbol grouping size (processing multiple binary symbols as one group) to reduce the frequency of coding operations. By adjusting the group size parameter, the system achieves a balance between compression ratio and computational load, reducing the number of probability updates and tree traversals required compared to traditional bit-by-bit CABAC.
2Measurement precision
If traditional entropy coding methods process binary sequences symbol by symbol, then coding accuracy is maintained, but processing speed decreases
Solution Approach 1:
The patent divides the binary sequence into multiple symbol groups and processes each group as a unified entity. This segmentation enables parallel processing potential and reduces the sequential operation count, thereby increasing processing speed while maintaining coding accuracy through context-adaptive probability modeling applied to each group.
Solution Approach 2:
The patent merges multiple binary symbols into a single group for joint processing. By combining multiple symbols and treating them as one coding unit, the system reduces the total number of coding operations required, improving processing throughput while preserving coding precision through context-adaptive probability estimation across the grouped symbols.
3Productivity
If multiple symbols are processed as a group, then processing throughput is improved, but memory requirements for code sets increase
Solution Approach 1:
The patent manages memory requirements by carefully selecting and managing the code sets for different symbol group sizes. Although larger groups require more comprehensive code sets, the use of context-adaptive coding allows the system to optimize memory usage by maintaining only the necessary code sets for current context states, balancing throughput improvement with acceptable memory consumption.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for encoding an input sequence of symbols to produce a bitstream and a method of decoding the bitstream to generate a reconstructed binary sequence. Encoding employs 2L-ary encoding trees having codewords associated with leaf nodes. A context model is used to select an encoding tree corresponding to an estimated probability at the encoder. The same context model is used by the decoder to select a decoding tree. The decoder interleaves bits from decoded sequences of length-L strings associated with different estimated probabilities, based on the context model.