Document Type Recognition Using Text and Structure Cues

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing OCR technologies struggle to accurately distinguish between similar document types such as invoices, delivery notes, order forms, quotations, and receipts using artificial intelligence, leading to errors in document type identification.

Innovation Solution

A document recognition apparatus and method that extracts text information, identifies item character strings and document structures, and determines document type based on a combination of these elements using a trained model and machine learning techniques.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If artificial intelligence models are used to determine document types, then processing speed is improved, but accuracy in distinguishing similar document types deteriorates

Engineering Contradiction:

Improveprocessing speedVSAvoiddocument type identification accuracy

Core Design Contradiction:

SpeedVSMeasurement precision

Solution Approach 1:

The patent segments the document type determination process into multiple stages: first extracting text information and identifying item character strings, then analyzing document structure, and finally combining both analyses for accurate classification. This segmentation allows the system to process documents efficiently while maintaining high accuracy by considering multiple features systematically.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent adds a structural dimension to the traditional text-based AI classification approach. By analyzing document structure (layout, formatting, positional relationships) in addition to text content, the system creates a multi-dimensional feature space that enables accurate distinction between similar document types while maintaining processing speed through efficient feature extraction and combination.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

2Device complexity

If only text information is extracted for document type determination, then processing complexity is reduced, but ability to distinguish similar document types deteriorates

Engineering Contradiction:

Improveprocessing complexityVSAvoiddocument type distinction capability

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The patent merges text information extraction with structural analysis into a unified document type determination system. The extracted text information (item character strings) and document structure are combined and processed together by the determination unit, creating a comprehensive analysis that accurately distinguishes similar document types without excessive processing complexity.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The patent creates a composite feature representation by combining text-based features (item character strings) with structure-based features (document layout, formatting characteristics). This composite approach leverages the strengths of both text analysis and structural analysis to achieve accurate document type classification while managing processing complexity through integrated processing.

Inventive Principle:

Principle #40Composite materials

3Measurement precision

If detailed structural analysis is performed, then document type classification accuracy is improved, but processing time increases

Engineering Contradiction:

Improvedocument type classification accuracyVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary extraction of text information and identification of item character strings before conducting structural analysis. This preliminary action prepares the data in advance, allowing the subsequent structural analysis to focus only on relevant features and reducing overall processing time while maintaining high classification accuracy through the combined analysis.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent implements a continuous processing flow where text extraction, item character string identification, and structural analysis are performed in an integrated manner. The determination unit continuously processes both text and structure features together, avoiding interruptions and redundant processing steps, thereby maintaining high accuracy while optimizing processing time through efficient continuous operation.

Inventive Principle:

Principle #20Continuity of useful action

Data Source

PatentUS20260011169A1Document recognition apparatus, document recognition method, and computer-readable, and non-transitory medium

Publication Date: 2026.01.08 PFU LTD

AI summary

A document recognition apparatus includes circuitry that extracts text information from document data, identifies an item character string and a structure of the document data from the extracted text information, the item character string being a character string for identifying a document type of the document data, and determines the document type of the document data, based on a combination of the item character string and the structure of the document data.