OCR Segmentation for Mixed Text Types Using Region-Specific Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing optical character recognition (OCR) technologies face challenges in effectively recognizing and processing multiple data/text types within a single document, such as printed text, curved text, and handwriting, due to the inability of single OCR models to handle diverse data types and lack of traceability of recognized text sources.
Innovation Solution
A system and method that detects different types of text data in a document, divides it into regions, removes optical noise using a self-attention mechanism, and selects the most suitable OCR software code for each region, enabling accurate extraction of computer-readable text through cognitive OCR modules, OCR recognition algorithm dispatching, and noise removal processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single OCR model is used to process the entire document, then the system complexity is low, but the recognition accuracy for diverse text types deteriorates
Solution Approach 1:
The patent divides the document into multiple text regions based on different text types (printed text, curved text, handwriting, tables, etc.) and applies different OCR models to each region. This segmentation allows each specialized model to optimize for its specific text type, improving overall recognition accuracy while managing complexity through modular architecture.
Solution Approach 2:
The patent implements local quality by tailoring the OCR processing approach to each specific text region. Different OCR models with specialized capabilities are selected and applied to different text types within the document, ensuring that each local region receives the most appropriate recognition technology for its characteristics.
2Measurement precision
If the document is divided into multiple text regions with different OCR models, then the recognition accuracy for diverse text types is improved, but the processing time increases
Solution Approach 1:
The patent performs preliminary classification of text regions before applying OCR models. By first identifying and categorizing different text types (printed, curved, handwriting, tables) and segmenting the document accordingly, the system prepares the structure in advance, allowing subsequent OCR processing to be more efficient and targeted rather than attempting universal processing.
3Measurement precision
If optical noise is not removed from text regions, then the processing speed is maintained, but the recognition accuracy deteriorates
Solution Approach 1:
The patent extracts and removes optical noise from text regions before applying OCR recognition. By separating the noise removal function from the main OCR processing pipeline and applying it selectively to each text region, the system improves recognition accuracy by eliminating interfering elements while maintaining efficient processing through targeted rather than universal noise removal.
Data Source
AI summary
A method, system, and computer program product for segmenting and processing documents for optical character recognition is provided. The method includes receiving a document and detecting different types of text data. The document is divided into a plurality of text regions associated with the different types of said text data. Optical noise is removed from each text region and differing optical character recognition software code is selected for application to each text region. The differing optical character recognition software code is executed with respect to each text region resulting in extractable computer readable text located within each said text region.