Character Identifying Subsystem for Grid-Based Text OCR
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Optical character recognition (OCR) applications often fail to accurately detect boundaries in scanned images, particularly for logographic characters like Japanese or Chinese, leading to errors such as broken characters or incorrectly merged characters, which can result in typographical and grammatical errors.
Innovation Solution
The proposed solution involves a character identifying subsystem that uses connected component analysis and grid layout analysis to improve OCR boundary detection. This subsystem processes scanned images by identifying connected elements as single characters, merging disjoint elements when necessary, and correcting errors through a second OCR pass with refined bounding information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional OCR is used for boundary detection, then the process is simple and fast, but accuracy deteriorates especially for logographic characters
Solution Approach 1:
The patent merges traditional OCR boundary detection with connected component analysis results. The system combines bounding boxes from OCR with bounding boxes from connected component analysis, merges overlapping boxes, and uses grid-based coordinate system to integrate results from multiple detection methods, thereby improving boundary detection accuracy for logographic characters while managing system complexity through structured integration
Solution Approach 2:
The patent introduces an intermediary processing layer that takes outputs from both OCR and connected component analysis, applies grid-based coordinate transformation, merges bounding boxes through systematic algorithms, and produces refined boundary information. This intermediary layer acts as a mediator that reconciles results from multiple sources without requiring complete system redesign
2Reliability
If connected component analysis is used alone, then character connectivity is preserved, but boundary precision deteriorates
Solution Approach 1:
The patent merges bounding boxes from connected component analysis with those from traditional OCR. By combining the results and merging overlapping bounding boxes using grid-based coordinate systems, the system preserves character connectivity information while refining boundary precision through the complementary strengths of both methods
Solution Approach 2:
The patent transforms coordinate systems to a grid-based framework, allowing precise representation and merging of bounding boxes from different sources. This parameter transformation enables systematic integration of connectivity-preserving boxes with precision-oriented boxes, resolving the contradiction between reliability and measurement precision
3Measurement precision
If multiple OCR passes are performed with refined boundaries, then accuracy improves, but processing time increases
Solution Approach 1:
The patent performs preliminary boundary refinement by merging OCR results with connected component analysis before the main OCR recognition process. By pre-processing and establishing accurate bounding boxes using grid-based coordination and merging algorithms, the system reduces the need for multiple corrective OCR passes, thereby improving accuracy while minimizing additional processing time
Data Source
AI summary
A method for identifying characters in scanned images of objects includes identifying a first set of characters in a scanned image of an object based on connected component analysis, identifying a second set of characters for the object based on an optical character recognition (OCR) process on the image of the object, and combining the first set of characters with the second set of characters to create a third set of characters.