Document Pixel Density Clustering for Non-Textual Data Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods are inadequate for accurately extracting non-textual data such as tables, flowcharts, and UML diagrams from electronic documents, as they often rely on location-dependent structures and require extensive tagging, making them inefficient in noisy environments.
Innovation Solution
A computer-implemented method that converts documents into pixels, creates a pixel density map, determines density clusters, calculates coordinates, identifies candidate information regions, and extracts textual data using image clustering and machine learning techniques, allowing for targeted extraction of useful regional areas without relying on structural location or extensive tagging.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing methods are used to extract non-textual data, then extraction can be performed, but the accuracy is inadequate and the methods rely on location-dependent structures
Solution Approach 1:
The patent transforms the document extraction problem from a structure-based approach to a density-based approach by converting the document into a pixel density map. This parameter transformation allows the system to identify information regions based on pixel density characteristics rather than fixed structural locations, thereby improving both extraction accuracy and location independence.
Solution Approach 2:
The patent replaces traditional mechanical/structural analysis methods with image processing and clustering algorithms. By treating the document as an image and applying density-based clustering, the system achieves more accurate extraction of non-textual data without relying on predetermined structural locations.
2Productivity
If traditional extraction methods are used, then data can be retrieved, but extensive tagging is required which reduces efficiency
Solution Approach 1:
The patent implements self-service by allowing the system to automatically identify information regions through density clustering without requiring manual tagging or annotation. The algorithm autonomously determines candidate information regions based on pixel density patterns, eliminating the need for extensive preprocessing tags and significantly improving extraction efficiency.
3Reliability
If conventional methods are applied, then extraction is performed, but noise reduction is insufficient in noisy environments
Solution Approach 1:
The patent applies local quality by analyzing pixel density characteristics in local regions rather than treating the entire document uniformly. By identifying density clusters and their coordinates, the system can selectively process candidate information regions with different density profiles, thereby improving noise resistance while preserving important information signals.
Data Source
AI summary
An approach for extracting non-textual data from an electronic document is disclosed. The approach includes receiving a request to extract a file and converting the file into pixels. The approach creates a pixel map of the converted file and determines one or more density clusters of the pixel map based on image clustering method. Furthermore, the approach determines one or more coordinates of the one or more density clusters and determines one or more candidate information regions based on the one or more coordinates, density of the one or more density clusters. Finally, the approach extracts one or more textual data based on the one or more candidate information regions and outputs the extracted one or more textual data.