Type-Aware Arithmetic Encoding to Cut Memory and Time Overhead

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing data compression methods using arithmetic encoding suffer from high memory and time overheads due to the use of a large number of data compression models without type-specific selection.

Innovation Solution

A computing device employs multiple processors to identify the data type of the input data and selectively applies data compression models tailored to that type, reducing the number of models used and offloading the compression process to a secondary processor, such as a coprocessor, to minimize memory and time overheads.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If a large quantity of data compression models are used for arithmetic encoding, then compression rate is improved, but memory overhead and encoding time increase

Engineering Contradiction:
Improvecompression rateVSAvoidmemory overhead
Core Design Contradiction:
ProductivityVSQuantity of substance

Solution Approach 1:

The patent segments the data compression process by dividing data into different types (e.g., text, image, audio, video) and selecting specific compression models for each type. This segmentation allows the system to use fewer models overall while maintaining high compression rates for each data category.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies local quality by using different compression models optimized for specific data types. Each data type receives a tailored compression approach (e.g., text models for text data, image models for images), ensuring optimal compression performance for each local context rather than using a universal model.

Inventive Principle:
Principle #3Local quality

2Productivity

If a large quantity of data compression models are used for arithmetic encoding, then compression rate is improved, but encoding time increases

Engineering Contradiction:
Improvecompression rateVSAvoidencoding time
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent segments the data compression process by dividing data into different types (e.g., text, image, audio, video) and selecting specific compression models for each type. This segmentation allows the system to use fewer models overall while maintaining high compression rates for each data category.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary action by classifying the data type before compression. This pre-classification step enables the system to select the appropriate compression model in advance, avoiding the time-consuming process of evaluating multiple models during the actual compression phase.

Inventive Principle:
Principle #10Preliminary action

3Productivity

If multiple processors are used to distribute compression workload, then compression performance is enhanced, but device complexity increases

Engineering Contradiction:
Improvecompression performanceVSAvoidprocessor architecture
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent implements a multi-processor architecture where each processor can handle different data types independently. The first processor handles data type identification and model selection, while the second processor executes the actual compression, allowing both processors to perform multiple functions and improving overall system efficiency.

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

Solution Approach 2:

The patent introduces an intermediary mechanism (the data type identification module and model selection logic) that coordinates between the first processor and second processor. This intermediary layer manages the workflow and data flow between processors, enabling efficient parallel processing while maintaining system coherence.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3944505B1Data compression method and computing device
Publication Date: 2025.10.22 HUAWEI TECH CO LTD
  • EP3944505B1 patent drawingFigure 1
  • EP3944505B1 patent drawingFigure 2~3
  • EP3944505B1 patent drawingFigure 4~5

AI summary

This application discloses a data compression method and a computing device, and relates to the data compression field, to resolve a problem of how to reduce memory overheads and time overheads for performing data compression through arithmetic encoding. The method is applied to the computing device. The method includes: receiving, by the computing device, to-be-compressed data, and identifying a data type of the to-be-compressed data; selecting one or more data compression models based on the data type; and compressing the to-be-compressed data based on the selected one or more data compression models.