OCR for Non-Coplanar Documents via Region-Specific Coordinate Transformations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing optical character recognition (OCR) methods face challenges in accurately recognizing text from images of documents with non-coplanar regions, such as bent or folded pages, due to inadequate coordinate transformations, leading to incorrect character recognition and degraded OCR quality.
Innovation Solution
The method involves identifying base points within each planar region of a document image, determining separate coordinate transformations for each region, and applying boundary corrections to improve the accuracy of coordinate transformations, allowing for the identification of symbol sequences across multiple images and producing a median symbol sequence for improved OCR text representation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single coordinate transformation is used for the entire document image, then the transformation process is simple, but OCR accuracy deteriorates for documents with non-coplanar regions
Solution Approach 1:
The patent divides the document image into multiple planar regions based on detected base points and applies separate coordinate transformations to each region. This segmentation allows each transformation to be optimized for its specific region's geometry, maintaining high OCR accuracy across non-coplanar surfaces while keeping individual transformations computationally manageable.
2Measurement precision
If separate coordinate transformations are determined for each planar region, then OCR accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent applies different coordinate transformation parameters to different planar regions of the document. Each region receives a transformation optimized for its specific geometric characteristics, allowing high precision OCR for each local area while the overall system manages complexity through region-based processing.
3Measurement precision
If base points are identified in multiple images, then symbol sequence matching accuracy is improved, but processing time increases
Solution Approach 1:
The patent performs preliminary identification of base points in multiple images before the main OCR processing. This preliminary action establishes accurate correspondence relationships between images in advance, enabling faster and more accurate symbol sequence matching during the actual recognition process without repeating computationally intensive operations.
Data Source
AI summary
Systems and methods for performing OCR of an image depicting text symbols and imaging a document having a plurality of planar regions are disclosed. An example method comprises: receiving a first image of a document having a plurality of planar regions and one or more second images of the document; identifying a plurality of coordinate transformations corresponding to each of the planar regions of the first image of the document; identifying, using the plurality of coordinate transformations, a cluster of symbol sequences of the text in the first image and in the one or more second images; and producing a resulting OCR text comprising a median symbol sequence for the cluster of symbol sequences.