Redaction Block Detection in Document OCR Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing document processing systems struggle to accurately detect and process redaction blocks in electronic documents, leading to incorrect identification and handling of redacted text, especially in automated systems, where OCR processes often return blank or garbage characters for redacted areas.
Innovation Solution
A method and system that transform documents into graphical or image files to identify redaction blocks and associated text fragments, using techniques like scanning for consecutive black pixels to detect redaction blocks and applying OCR to locate text fragments, with reconciliation to map and correct the representation of redacted text, preventing OCR from misinterpreting it as missing data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing OCR processes are used to scan documents, then text extraction is performed, but redacted text blocks return blank or garbage characters instead of accurate detection
Solution Approach 1:
The patent segments the document analysis into distinct phases: first identifying redaction blocks through image processing (detecting blacked-out areas), then performing OCR on non-redacted regions separately. This segmentation prevents OCR from misinterpreting redacted areas as text, resolving the contradiction between text extraction and redacted text identification accuracy.
Solution Approach 2:
The patent performs preliminary identification of redaction blocks before applying OCR processing. By pre-marking redacted areas and excluding them from OCR analysis, the system ensures that OCR only processes visible text, thereby preventing garbage character generation and improving both detection precision and reliability.
2Extent of automation
If manual visual inspection is used to verify text and image connection, then accuracy can be achieved, but automated processing cannot be performed
Solution Approach 1:
The patent introduces an intermediary layer of redaction block detection that acts as a mediator between the image data and OCR processing. This intermediary identifies and marks redacted areas, enabling automated systems to make decisions about which regions require manual review versus which can be processed automatically, thus maintaining both automation and accuracy.
Solution Approach 2:
The patent implements a feedback mechanism where the system continuously refines its redaction block detection based on OCR results and image analysis. By comparing detected redaction blocks with OCR output and adjusting detection parameters accordingly, the system achieves high accuracy in automated processing without requiring manual verification of every text instance.
3Productivity
If redaction blocks are not accurately detected, then processing speed is maintained, but correct categorization and handling of redactions fails
Solution Approach 1:
The patent replaces manual mechanical review processes with automated image processing and pattern recognition algorithms. By using computational methods to detect redaction blocks through pixel analysis and pattern matching, the system achieves both high processing speed and reliable redaction identification, eliminating the need for slow manual verification while maintaining accuracy.
Data Source
AI summary
A computer-implemented method, system and a computer program product are provided for automatically detecting redaction blocks in an image file document by analyzing the document to identify any redaction block areas and then detecting location information for each redaction block area identified in the document which may be mapped to any associated text fragments in the document based on the location information for each redaction block area and text fragment in the document.