OCR Candidate String Selection for Faster Document Digitization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing optical character recognition (OCR) systems require significant user effort for manual correction due to misrecognitions, and existing methods do not effectively utilize candidate character strings from other regions of an image for input assistance.
Innovation Solution
An information processing system that includes a multifunction peripheral (MFP) and external storage, which extracts candidate character strings from OCR results and displays them for user selection, reducing the need for manual input by allowing users to choose from suggested character strings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If OCR is used for data entry, then the speed of digitizing documents is improved, but the accuracy of character recognition deteriorates due to misrecognitions
Solution Approach 1:
The patent introduces an intermediary character selection interface that mediates between OCR results and final text output. When OCR accuracy is insufficient, the system displays candidate characters from the image for user selection, allowing the workflow to continue at high speed while correcting accuracy issues through selective user input rather than complete manual retyping
Solution Approach 2:
The system enables self-service by automatically generating candidate character lists from the image data and presenting them for user confirmation. This allows users to quickly correct OCR errors by selecting from pre-generated candidates rather than manually typing entire text passages, maintaining high digitization speed while improving accuracy
2Measurement precision
If manual correction of OCR results is performed, then the accuracy of character recognition is improved, but the time required for processing deteriorates
Solution Approach 1:
Instead of requiring complete manual verification of all OCR results, the system applies partial action by only prompting users to select characters when OCR confidence is insufficient. This selective approach maintains high accuracy for confident recognitions while minimizing time loss for corrections only where needed
Solution Approach 2:
The system performs preliminary action by pre-generating candidate character lists from the image data before user interaction. This preparation work is done automatically, so when users need to correct errors, they only need to select from pre-computed candidates rather than generating corrections from scratch, significantly reducing correction time
3Ease of operation
If candidate character strings are displayed for user selection, then the ease of operation is improved, but the device complexity deteriorates
Solution Approach 1:
The system achieves universality by making the MFP perform multiple functions: it not only scans and performs OCR but also generates candidate character lists, manages display interfaces, and handles character selection logic. This consolidates what could be separate complex systems into a single multi-functional device, improving ease of operation without proportionally increasing overall system complexity
Data Source
AI summary
Multiple character strings included in data are obtained. An input of a character is received from a user. One or more character strings corresponding to the input character are displayed on a display unit. A character string to be saved in association with a file generated based on the data is set by using a character string selected from the one or more character strings displayed.