Probability Interval Partitioning for Adaptive Entropy Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing entropy coding methods, such as Huffman codes, face challenges in adapting to time-varying source statistics and dealing with highly skewed probability distributions, leading to inefficiencies and increased complexity, especially when handling higher-order probability modeling and dominant alphabet values.

Innovation Solution

The proposed method employs a binary arithmetic coding engine that maps source symbols into binary decisions, using binarization schemes and parameter assignment to create partial bitstreams, allowing for efficient encoding and decoding by decoupling probability modeling from entropy coding and enabling adaptive coding with reduced complexity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If Huffman codes with VLC tables are used for entropy coding, then the implementation is simple and efficient for fixed probability distributions, but the adaptation to time-varying source statistics becomes complex and costly

Engineering Contradiction:
Improveimplementation simplicityVSAvoidadaptation to time-varying statistics
Core Design Contradiction:
Ease of manufactureVSAdaptability or versatility

Solution Approach 1:

The patent implements dynamic adaptation of VLC tables based on time-varying source statistics. The system monitors symbol frequencies and automatically updates the VLC table structure to match changing probability distributions, allowing the coding scheme to adapt dynamically without requiring complex reconfiguration procedures.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes the parameters of the VLC tables (code assignments, table structure) based on observed source statistics. By adjusting the VLC table parameters in response to changing symbol probabilities, the system maintains coding efficiency while adapting to time-varying characteristics of the data source.

Inventive Principle:
Principle #35Parameter changes

2Device complexity

If Huffman codes are used without alphabet extension, then the implementation remains simple, but the redundancy increases substantially for dominant alphabet values

Engineering Contradiction:
Improvecoding scheme complexityVSAvoidcoding redundancy
Core Design Contradiction:
Device complexityVSLoss of information

Solution Approach 1:

The patent segments the alphabet into multiple categories or groups, allowing different coding strategies to be applied to different segments. By dividing the alphabet structure, the system can handle dominant values more efficiently through specialized coding techniques while maintaining overall simplicity.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces asymmetric coding treatments for different symbol types, particularly for dominant alphabet values. By applying asymmetric extensions or specialized coding rules for frequent symbols, the system reduces redundancy for dominant values without significantly increasing overall complexity.

Inventive Principle:
Principle #4Asymmetry

3Measurement precision

If multiple sets of VLC tables are used for higher-order probability modeling, then the coding accuracy improves, but the implementation complexity and memory requirements increase

Engineering Contradiction:
Improveprobability modeling accuracyVSAvoidnumber of VLC tables
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent designs a universal VLC table structure that can serve multiple probability modeling orders. By creating a multi-functional coding table that adapts to different context lengths and probability models, the system achieves higher-order modeling accuracy without requiring separate dedicated tables for each modeling level.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent merges multiple VLC tables into a unified structure or combines them in a hierarchical manner. By merging the functionality of multiple tables into a single adaptive structure, the system reduces memory requirements and access complexity while maintaining the precision benefits of higher-order probability modeling.

Inventive Principle:
Principle #5Merging (Combining)

4Productivity

If arithmetic coding is used for handling skewed probability distributions, then the coding efficiency improves, but the implementation complexity increases substantially compared to VLC

Engineering Contradiction:
Improvecoding efficiencyVSAvoidalgorithmic complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent employs simplified, approximate arithmetic coding techniques that use pre-computed lookup tables and fixed-precision arithmetic. By using disposable, pre-calculated probability models and avoiding complex iterative calculations, the system achieves arithmetic coding efficiency with reduced implementation complexity.

Inventive Principle:
Principle #27Cheap short-living objects (Disposable)

Solution Approach 2:

The patent performs preliminary computation of probability models and coding parameters before the actual encoding process. By pre-calculating probability distributions, code intervals, and lookup tables in advance, the system reduces the complexity of the real-time encoding operation while maintaining the efficiency benefits of arithmetic coding for skewed distributions.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP2559166B1Probability interval partioning encoder and decoder
Publication Date: 2017.11.01 FRAUNHOFER GESELLSCHAFT ZUR FORDERUNG DER ANGEWANDTEN FORSCHUNG EV
  • EP2559166B1 patent drawingFigure 1
  • EP2559166B1 patent drawingFigure 2
  • EP2559166B1 patent drawingFigure 3~4

AI summary

An encoder for encoding a sequence of symbols is described which comprises an assigner configured to assign a number of parameters to each symbol of the sequence of symbols based on information contained within previous symbols of the sequence of symbols; a plurality of entropy encoders each of which is configured to convert the symbols forwarded to the respective entropy encoder into a respective bitstream; and a selector configured to forward each symbol to a selected one of the plurality of entropy encoders, the selection depending on the number of parameters assigned to the respective symbol.