Document Classification via Geometric Object Positioning

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing document classification methods, such as OCR and human vision imitation, face challenges with tilting, displacement, and scaling issues during digitization, requiring significant hardware resources and complex operations.

Innovation Solution

A system comprising a database, positioning-information capturing device, locating device, and document comparison device that captures and compares positioning information to align and orient digitized documents with reference documents, eliminating interference factors like tilting, displacement, and scaling through rotational, displacement, and scaling treatments.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If OCR is used for automatic document classification, then classification accuracy is improved, but hardware resource requirements and operational complexity increase significantly

Engineering Contradiction:

Improveclassification accuracyVSAvoidhardware resource requirements

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts only the essential positioning information (coordinates, distances, angles) of objects in documents, separating this from full text recognition. This extraction approach achieves classification accuracy without requiring complete OCR processing, thereby reducing hardware resource requirements and operational complexity.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent replaces the mechanical OCR process with a geometric comparison system that uses coordinate data, distance calculations, and angle measurements. This substitution eliminates the need for complex text recognition hardware while maintaining classification effectiveness through mathematical relationships between document objects.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Adaptability or versatility

If human vision imitation method is used to capture document characteristics, then classification capability is improved, but interference from tilting, displacement, and scaling affects accuracy

Engineering Contradiction:

Improveclassification capabilityVSAvoidclassification accuracy

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent changes the parameter representation from absolute coordinates to relative geometric relationships (distances and angles between objects). This parameter transformation makes the classification system invariant to tilting, displacement, and scaling, as these transformations preserve relative geometric relationships while changing absolute positions.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent creates an equipotential classification system where all documents are evaluated based on their relative geometric relationships rather than absolute positioning. This approach equalizes the classification potential across documents with different orientations, positions, or scales, eliminating the interference these factors cause to accuracy.

Inventive Principle:

Principle #12Equipotentiality

3Measurement precision

If comprehensive document processing is performed to eliminate tilting, displacement, and scaling, then classification accuracy is improved, but operational complexity and processing time increase

Engineering Contradiction:

Improveclassification accuracyVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent extracts only the essential geometric relationships (distances and angles between objects) needed for classification, ignoring unnecessary details about absolute positions and orientations. This selective extraction achieves accurate classification without requiring time-consuming processing to correct tilting, displacement, or scaling issues.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies partial action by performing only the necessary geometric comparisons rather than comprehensive document processing. By calculating only the essential distances and angles between key objects, the system achieves classification accuracy without the excessive processing time required for full document normalization and correction.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS7602971B2System, method and recording medium for automatically classifying documents

Publication Date: 2009.10.13 NEWSOFT TECH CORP

AI summary

A system for automatically classifying documents includes a database, a positioning-information capturing device, a locating device, and a document comparison device. The database is used for storing a reference document. The positioning-information capturing device is used for capturing the positioning information of a plurality of objects in a digitized document. The locating device is used for capturing the corresponding objects of the reference document according to the positioning information of the objects. The document comparison device then proceeds with an orientation process and determines whether the digitized document and the reference document are of the same classification according to the distances between all of the objects of the digitized document and the nearest objects thereto in the reference document. A method and a computer readable recording medium for automatically classifying documents are also disclosed.