Binary Entropy Coding with Reduced Probability Sources
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing binary entropy coding methods face inefficiencies due to overhead in parallel encoding and two-codeset variable-to-variable encoding, particularly with a large number of sources, leading to increased computational demands and secondary codeword usage.
Innovation Solution
Selecting a new probability set for entropy encoding and decoding, with a mapping that reduces the number of probabilities, thereby minimizing overhead and secondary codewords, using a method that optimizes relative entropy mapping to improve encoding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If parallel encoding with multiple sources is used, then coding efficiency is improved, but overhead information about the length of each subsequence increases
Solution Approach 1:
The patent merges multiple probability sources into a reduced set of probability sources. Instead of maintaining separate encoding streams for each of the 64 probability sources, the invention combines them into a smaller number of merged sources, thereby reducing the overhead information required to describe each subsequence while preserving the parallel encoding efficiency benefits.
Solution Approach 2:
The patent creates a universal mapping framework that can handle any of the 64 probability sources through a reduced set of merged probability sources. This universal approach allows the same reduced probability set to serve multiple original sources, reducing overhead while maintaining the ability to encode diverse probability distributions efficiently.
2Measurement precision
If a large number of probability sources are used, then coding precision is improved, but the number of secondary codewords increases
Solution Approach 1:
The patent combines multiple probability sources into a reduced set of merged probability sources. By merging sources that have similar statistical characteristics, the invention maintains coding precision for the overall data while reducing the number of secondary codewords needed to handle end-of-subsequence events.
Solution Approach 2:
The patent changes the parameter of probability source granularity by transitioning from 64 fine-grained probability sources to a smaller number of merged probability sources. This parameter change reduces the complexity of managing secondary codewords while preserving essential coding precision through the mapping relationship.
3Productivity
If CABAC is used instead of CAVLC, then compression ratio is improved, but computational complexity increases
Solution Approach 1:
The patent merges multiple probability computation paths into a reduced set, thereby reducing the computational complexity of CABAC while maintaining its superior compression ratio advantages over CAVLC. The merging approach reduces the number of context model updates and probability calculations required.
Solution Approach 2:
The patent changes the computational parameters by reducing the number of active probability sources from 64 to a smaller merged set. This parameter reduction decreases the computational burden of arithmetic coding operations while preserving the adaptive compression benefits that give CABAC its superior compression ratio.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Methods and systems are described for combining sources in a data compression system. In a system in which a context model results in the production of multiple sources or symbols, each source being associated with a probability estimate, sources may be combined by defining another, reduced size, set of probabilities. The new set of probabilities may be a subset of the predefined set of probabilities provided by the context model. Minimizing relative entropy may be a basis upon which to define a mapping of predefined probabilities to the new set of probabilities. An input sequence that was modeled using the context model may then be entropy encoded and entropy decoded using the new set of probabilities based on a mapping between the new probabilities and the predefined probabilities.