Document Digitization Page Verification with OCR Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current document digitization processes, involving scanning and optical character recognition (OCR), lack a fast and robust verification methodology to ensure complete data capture, often missing handwritten remarks or marginal information.
Innovation Solution
A page verifier stage is added after OCR, allowing operators to rapidly review and identify missing areas by removing or highlighting recognized objects, enabling simultaneous viewing of multiple pages and automatic detection of neglected content for further processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If operators review the entire page to be entered, then they can see all information including handwritten remarks, but the process becomes laborious and productivity decreases
Solution Approach 1:
The verification interface is segmented into distinct regions: the original scan area and the recognized content area. This segmentation allows operators to focus on specific areas of interest rather than viewing the entire page, thereby maintaining productivity while ensuring completeness of data capture.
Solution Approach 2:
The patent introduces an intermediary visual representation that bridges the original scan and the recognized content. This intermediary interface presents information in a structured manner, allowing operators to efficiently verify completeness without laboriously reviewing the entire page, thus resolving the contradiction between reliability and productivity.
2Productivity
If operators see only the word being corrected or small snippets, then productivity is enhanced, but information may be missed that was omitted by OCR
Solution Approach 1:
The patent adds a spatial dimension to the verification process by presenting the original scan and recognized content in separate visual regions. This dimensional arrangement allows operators to quickly scan for omissions while maintaining focus on specific words or snippets, thereby enhancing both productivity and reliability simultaneously.
Solution Approach 2:
The verification interface is segmented into distinct regions: the original scan area and the recognized content area. This segmentation allows operators to focus on specific areas of interest rather than viewing the entire page, thereby maintaining productivity while ensuring completeness of data capture.
3Reliability
If manual verification is performed on the entire document, then completeness is ensured, but the verification time increases significantly
Solution Approach 1:
The system performs preliminary OCR recognition and prepares the recognized content for verification before the operator begins manual review. This preliminary action reduces the verification time by having the recognized content ready in an organized format, allowing operators to quickly compare and identify omissions without performing full manual verification of the entire document.
Solution Approach 2:
The patent adds a spatial dimension to the verification process by presenting the original scan and recognized content in separate visual regions. This dimensional arrangement allows operators to quickly scan for omissions while maintaining focus on specific words or snippets, thereby enhancing both productivity and reliability simultaneously.
Data Source
AI summary
Techniques for performing page verification of a document are provided. The techniques include performing a recognition technique on a document to recognize one or more objects in the document, excluding the one or more recognized objects from the document, and performing page verification of the document, wherein page verification comprises visual inspection of the document excluding the one or more recognized objects.