Character Encoding Detection via Statistical Fingerprinting

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for identifying character encodings, particularly for the ISO 8859 series, are prone to misidentification due to similarities in encoding schemes, leading to high chances of error when distinguishing between them.

Innovation Solution

A statistical classification method is employed to generate a fingerprint from data analysis results, using confidence values and frequency values of byte occurrences to accurately determine character encodings, which reduces incorrect determinations by considering all data analysis results rather than relying on dominant outcomes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional confidence-based methods are used to identify character encodings, then the identification process is simple and fast, but the accuracy is low especially for ISO 8859 series encodings

Engineering Contradiction:
Improvecharacter encoding identification accuracyVSAvoididentification system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent combines multiple data analysis results (confidence values from encoding detection, frequency values from byte analysis, and statistical features) into a unified fingerprint representation. This merging of multiple indicators resolves the contradiction by achieving high accuracy through comprehensive analysis while maintaining systematic processing efficiency.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent transforms the identification approach by changing from relying solely on confidence values to using a multi-parameter fingerprint including confidence values, byte frequency distributions, and statistical features. This parameter expansion enables accurate distinction between similar ISO 8859 encodings without excessive system complexity.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If statistical classification with comprehensive data analysis is used, then the accuracy of character encoding detection is improved, but the processing time and computational complexity increase

Engineering Contradiction:
Improvecharacter encoding detection accuracyVSAvoiddetection processing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary data analysis by pre-computing confidence values for multiple encodings and byte frequency distributions before classification. This preliminary action stores essential features in advance, enabling fast statistical classification without re-processing raw data, thus resolving the time-accuracy contradiction.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent replaces traditional sequential analysis methods with statistical classification that processes multiple features simultaneously. This substitution enables parallel computation of confidence values and frequency analyses, reducing processing time while maintaining comprehensive accuracy.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP2506154B1Text, character encoding and language recognition
Publication Date: 2014.07.09 CLEARSWIFT
  • EP2506154B1 patent drawingFigure 1
  • EP2506154B1 patent drawingFigure 2~3
  • EP2506154B1 patent drawingFigure 4~5

AI summary

A method is disclosed, for recognizing whether some electronic data is the digital representation of a piece of text and, if so, in which character encoding it has been encoded. A fingerprint is constructed from the data, wherein the fingerprint comprises, for each of a plurality of predetermined character encoding schemes, at least one confidence value, representing a confidence that the data was encoded using said character encoding scheme. The fingerprint also comprises a frequency value for each of a subset of byte values, each frequency value representing the frequency of occurrence of a respective byte value in the data. A statistical classification of the data is then performed based on the fingerprint. The method may be applied to spam classification.