Document Classification Warning System for Similarity Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing document classification methods often fail to accurately classify documents due to similarities in feature settings, leading to misclassification, especially when multiple similar features are used.

Innovation Solution

An information processing apparatus that receives document data, extracts feature data, compares it with existing data, and outputs a warning when similarity exceeds a threshold, allowing users to register definition data with high accuracy by distinguishing between similar documents through threshold value determination and other classification processes like patch, text recognition, or code processing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If multiple similar features are set for document classification, then the classification system can handle more document types, but the accuracy of classification decreases due to misclassification

Engineering Contradiction:

Improvedocument type coverageVSAvoidclassification accuracy

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The system performs similarity comparison between the input document and registered definition data, then provides feedback in the form of a warning message to the user. This feedback mechanism allows the user to be aware of potential misclassification risks and adjust the classification settings accordingly, thus maintaining high accuracy even when multiple similar features are used

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The similarity comparison is performed in advance before the final classification is made. By checking for similarity between the input document and existing definition data beforehand, the system can warn the user about potential classification conflicts, allowing preliminary adjustment of classification parameters to avoid misclassification

Inventive Principle:

Principle #10Preliminary action

2Productivity

If definition data is registered without checking for similarity with existing data, then the registration process is faster, but the classification accuracy decreases due to duplicate or similar entries

Engineering Contradiction:

Improveregistration speedVSAvoidclassification accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The system performs a partial check by comparing only key features of the input document with registered definition data, rather than a complete analysis. This partial similarity comparison is sufficient to detect obvious duplicates or highly similar entries, maintaining registration speed while preventing inaccurate classification

Inventive Principle:

Principle #16Partial or excessive action

3Measurement precision

If the system performs similarity comparison for every document registration, then the classification accuracy is improved, but the processing time increases

Engineering Contradiction:

Improveclassification accuracyVSAvoidregistration time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system changes the parameter of similarity comparison by adjusting the threshold value. By setting an appropriate threshold, the system can quickly identify documents that are sufficiently similar to trigger a warning, while allowing documents below the threshold to be registered without delay. This parameter adjustment balances accuracy improvement with time efficiency

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS11354496B2Information processing apparatus and non-transitory computer readable medium storing program

Publication Date: 2022.06.07 FUJIFILM BUSINESS INNOVATION CORP

AI summary

An information processing apparatus includes a processor configured to: receive document data representing a document in a case of registering definition data which defines classification of the document, the document data being data for creating the definition data; and output a warning in a case where the document data is similar to another document data of another definition data already registered.