ML System for Auto Tagging Unstructured PDF Accessibility

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Many PDF documents lack accessibility tags, making them inaccessible to visually impaired individuals, and existing technologies struggle to automatically add these tags to unstructured PDF files.

Innovation Solution

A machine learning system that processes unstructured PDF files by identifying graphical objects and associating them with Document Object Model (DOM) structure elements, generating training records, and using these records to train a machine learning model, which then automatically tags unstructured PDF documents to meet accessibility standards like PDF/UA, WCAG, and HHS Section 508.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If manual tagging of PDF documents is performed to ensure accessibility compliance, then accessibility quality is improved, but time consumption and labor costs increase significantly

Engineering Contradiction:

Improveaccessibility complianceVSAvoidtagging time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system enables documents to automatically generate their own accessibility tags through machine learning analysis. The ML model processes document structure, headings, paragraphs, tables, and images to autonomously assign appropriate accessibility tags without requiring manual human intervention, thus resolving the contradiction between compliance quality and time consumption

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces the manual mechanical process of accessibility tagging with an automated machine learning system. The ML model analyzes document features and generates accessibility tags algorithmically, substituting human labor with computational processing to achieve both high compliance quality and efficient processing speed

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Productivity

If automated tagging systems are implemented for unstructured PDF files, then processing speed is improved, but tagging accuracy and compliance quality deteriorate

Engineering Contradiction:

Improveprocessing speedVSAvoidtagging accuracy

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The system performs preliminary analysis of document structure, identifying headings, paragraphs, tables, images, and other structural elements before generating accessibility tags. This preparatory processing enables the ML model to make accurate tagging decisions based on comprehensive structural understanding, maintaining high accuracy while achieving automated processing speed

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The ML model incorporates feedback mechanisms where it analyzes the relationship between document structural features and accessibility requirements. The system learns from the document structure patterns and adjusts tagging decisions based on identified relationships, improving accuracy while maintaining automated processing efficiency

Inventive Principle:

Principle #23Feedback

3Reliability

If comprehensive accessibility tags are added to PDF documents, then accessibility quality is improved, but device complexity and processing requirements increase

Engineering Contradiction:

Improveaccessibility qualityVSAvoidsystem complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system segments the accessibility tagging process into distinct analytical components: identifying document structure elements (headings, paragraphs, tables, images), analyzing relationships between elements, and generating appropriate accessibility tags for each segment. This segmentation enables comprehensive accessibility tagging while managing system complexity through modular processing

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS12118294B2Machine learning systems and methods for automatically tagging documents to enable accessibility to impaired individuals

Publication Date: 2024.10.15 OPEN TEXT CORP

AI summary

Systems, methods, and products for auto tagging structured PDF documents that do not have accessibility tags. In one embodiment, structured PDF documents having accessibility tags are first parsed and analyzed to organize the visual components of the documents. The relationships of the identified objects to DOM elements (e.g., tags) are determined, and the objects and related DOM elements are stored in training files. The training files are used to train various classifiers. Untagged PDF documents are then parsed to identify included visual objects, and the classifiers are used to determine DOM elements that should be associated with visual objects identified in the untagged PDF documents. This information is used to construct a DOM structure corresponding to each untagged document. A new PDF is then generated corresponding to each untagged document using the generated DOM structure and visual object information.