Text Extraction Using Spatial and Contextual Relationships

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional text extraction techniques for documents are inflexible and resource-intensive, struggling with changes in content and layout, requiring frequent updates of regular expressions and templates to accurately identify textual elements.

Innovation Solution

A computing device uses spatial factors and contextual relationships between portions of computer-readable text, employing a computer-implemented model to identify and extract textual elements, reducing the need for frequent updates and resource usage.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If conventional text extraction techniques using regular expressions and templates are used, then textual elements can be identified and extracted from documents, but the system becomes inflexible when document content or layout changes, requiring frequent updates of regular expressions and templates

Engineering Contradiction:

Improveadaptability to document changesVSAvoidcomplexity of regular expressions and templates

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent replaces the mechanical system of regular expressions and templates with a machine learning-based natural language processing system. The NLP model automatically adapts to document changes by learning from document structures and content patterns, eliminating the need for manual updates of extraction rules when documents evolve.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system changes from fixed parameter matching (regular expressions) and position-based templates to dynamic parameter learning through machine learning. The NLP model learns optimal extraction parameters from training data and automatically adjusts to document variations without requiring explicit parameter updates.

Inventive Principle:

Principle #35Parameter changes

2Productivity

If conventional text extraction techniques are used, then textual elements can be extracted, but computing resources are consumed due to the need for frequent updates and maintenance of extraction rules

Engineering Contradiction:

Improveextraction efficiencyVSAvoidcomputing resource consumption

Core Design Contradiction:

ProductivityVSUse of energy by moving object

Solution Approach 1:

The system performs preliminary action by training the NLP model in advance on representative document data. This pre-training enables the model to handle document variations without requiring real-time computation for rule updates, reducing ongoing computing resource consumption during actual extraction operations.

Inventive Principle:

Principle #10Preliminary action

3Measurement precision

If regular expressions are updated to accommodate content changes, then extraction accuracy is maintained, but the burden on computing device resources increases

Engineering Contradiction:

Improveextraction accuracyVSAvoidtime for updating regular expressions

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The NLP-based system provides self-service by automatically adapting to document content changes without human intervention. The model continuously learns from document patterns and adjusts its extraction behavior autonomously, eliminating the time burden of manual regular expression updates while maintaining high extraction accuracy.

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS20250209846A1Computing system for extraction of textual elements from a document

Publication Date: 2025.06.26 HYLAND SOFTWARE INC

AI summary

Described herein are various technologies pertaining to text extraction from a document. A computing device receives the document. The document comprises computer-readable text and a layout, wherein the layout defines positions of the computer-readable text. Responsive to receiving the document, the computing device identifies at least one textual element in the computer-readable text based upon spatial factors between portions of the computer-readable text and contextual relationships between the portions of the computer-readable text. The computing device then outputs the at least one textual element.