Document Grouping via Marked Area Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current OCR techniques fail to effectively process printed forms with watermarks, impressions, or low-resolution images, leading to costly manual labor in back offices.
Innovation Solution
A method and system that mark specific areas in electronic documents, match them across multiple documents, compute scores based on matching, and group documents using a processor, reducing manual effort by employing pattern matching and scoring modules.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If OCR techniques are used to process printed forms, then digitization of textual fields is achieved, but processing fails when watermarks, impressions, or low-resolution images are present
Solution Approach 1:
The patent segments the document processing into multiple stages: initial OCR attempt, failure detection, template generation from successful cases, and template-based matching for failed cases. This segmentation allows the system to handle different document types and quality levels appropriately, maintaining high automation while improving reliability through fallback mechanisms.
Solution Approach 2:
The patent creates templates from successfully processed documents and uses these templates to guide processing of similar documents. By copying the structural patterns from good-quality documents, the system can reliably process documents with watermarks or low resolution, as the template provides a reference framework that compensates for poor image quality.
2Reliability
If manual labor is used to process documents with OCR failures, then processing accuracy is maintained, but time and money costs increase
Solution Approach 1:
The system performs self-service by automatically detecting OCR failures and invoking template-based processing without human intervention. The automated workflow identifies which documents need special handling and processes them through the appropriate pipeline, maintaining accuracy while preserving productivity by eliminating manual intervention for routine cases.
Solution Approach 2:
The system implements feedback loops where processing results are analyzed to identify failures, generate new templates from successful cases, and update the template library. This continuous learning process improves accuracy over time while maintaining automated processing, as the system adapts to new document types and quality variations without requiring manual reconfiguration.
3Productivity
If template-based processing is implemented, then processing speed and automation are improved, but system complexity increases
Solution Approach 1:
The system performs preliminary actions by pre-processing documents to extract templates before the main processing batch. Templates are generated in advance from representative samples and stored for rapid matching during production processing. This preliminary preparation simplifies the main processing pipeline, as documents can be quickly matched against existing templates without requiring complex real-time analysis.
Solution Approach 2:
The template library is designed to be dynamic and adaptive, automatically updated with new templates from successfully processed documents. The system dynamically selects which template to use based on document characteristics, and the template collection evolves over time to cover more document variations. This dynamic approach maintains simplicity by using straightforward matching logic while handling increasing document diversity.
Data Source
AI summary
A method, a system, and a computer program product for processing a plurality of electronic documents are provided. The method includes marking at least one area in an electronic document selected from the plurality of electronic documents. The marked at least one area is then matched with corresponding area in the remaining electronic documents of the plurality of electronic documents. A score is computed for each of the remaining electronic documents of the plurality of electronic documents based on the matching. The plurality of electronic documents is then grouped into one or more groups based on the score.