Region-Based OCR Accuracy Heat Map for Correction Guidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing optical character recognition (OCR) systems face challenges in accurately identifying erroneous recognition cases beyond character combinations, such as color characters, slanted text, and image quality issues, leading to inadequate heat map displays for guiding user corrections.
Innovation Solution
A system that records correction history information for each region of an original image, calculates recognition accuracy based on this data, and generates a distribution image displaying accuracy differences, allowing for targeted attention in correction work across various error-prone regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If heat map display is performed based on recognition accuracy parameters for individual words, then users can identify problematic regions, but the system cannot account for various error causes such as color characters, slanted text, and image quality issues
Solution Approach 1:
The patent segments the original image into multiple regions and performs separate OCR processing for each region. Correction history is recorded per region, allowing the system to track errors specific to different areas. This segmentation enables the heat map to reflect region-specific accuracy rather than word-level accuracy, capturing systematic errors in particular regions caused by factors like slanted text or image quality issues.
Solution Approach 2:
The patent transitions from one-dimensional word-level accuracy measurement to two-dimensional region-based accuracy measurement. By organizing correction history spatially across regions and generating a heat map that visualizes accuracy distribution across the image space, the system adds a spatial dimension to accuracy assessment, enabling better identification of region-specific error patterns.
2Ease of operation
If region-based correction history is recorded and accuracy is calculated for each region, then distribution image can show accuracy differences across regions, but the system complexity increases
Solution Approach 1:
The patent creates a multi-functional system where the same OCR and correction infrastructure serves both word-level and region-level accuracy assessment. The heat map generation unit reuses existing correction history data and overlays it on region boundaries, allowing one system to provide both traditional word-level correction support and new region-level guidance without duplicating core OCR functionality.
Solution Approach 2:
The patent introduces a heat map generation unit as an intermediary layer between the OCR processing system and the user interface. This intermediary takes correction history data and original image regions as input, generates visual heat map guidance, and presents it to users without requiring fundamental changes to the underlying OCR or correction systems.
3Reliability
If correction accuracy is tracked for each region instead of individual words, then systematic errors in specific regions can be identified, but the granularity of error detection is reduced
Solution Approach 1:
The patent applies local quality by assigning different accuracy characteristics to different regions based on their correction history. Each region develops its own accuracy profile reflecting systematic error patterns specific to that area. This allows the heat map to provide locally optimized guidance, where users can trust the accuracy assessment for each region based on its own correction experience rather than a global average.
Data Source
AI summary
A correction history recording unit that records region information of a correction site with respect to text data converted from an original image as correction history information, an accuracy calculation unit that calculates accuracy of optical character recognition for each of individual regions on a layout of the original image on the basis of the correction history information, a distribution image generation unit and a distribution image display unit which generate and display a distribution image in which a difference in magnitude of accuracy is shown as a difference in a display aspect for every individual region are included so as to generate and display the distribution image that is distinguished for every individual region by reflecting a tendency in which a character recognition rate in a certain region on a layout of the original image may decrease due to various cases including a format of an original document, a state of an OCR device, and the like.