Text Line Curvature Correction for Optical Character Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Text-containing images captured by hand-held devices or dedicated scanners often suffer from noise, optical blur, and curvature distortions due to curved-page-surface-induced and perspective-induced effects, which degrade the performance of optical-character recognition systems.
Innovation Solution
A method and system that identify the outline of a text-containing page, generate contours for text lines, determine centroids and inclination angles, and use a model to assign local displacements to pixels, thereby straightening text lines through an inclination-angle map or pixel-displacement map to correct for curvature.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If hand-held devices or dedicated scanners are used to capture text-containing images, then portability and ease of use are improved, but curvature distortions and perspective-induced effects worsen the quality of text images
Solution Approach 1:
The patent applies preliminary image processing operations including noise reduction, optical deblurring, and curvature correction before optical character recognition. By performing these corrective actions in advance, the system compensates for the distortions introduced by hand-held device portability, thereby maintaining both ease of operation and image quality.
2Adaptability or versatility
If curved-page-surface-induced and perspective-induced curvature effects occur during image capture, then portability and flexibility are improved, but text line straightness and character recognition accuracy worsen
Solution Approach 1:
The patent transforms the curved text lines into straight lines by applying geometric transformations and parameter adjustments. The system detects curvature parameters and applies corresponding correction transformations to straighten text lines, thereby maintaining adaptability to different capture conditions while improving character recognition accuracy.
3Ease of operation
If noise and optical blur are present in captured images, then ease of capture is improved, but optical character recognition performance worsens
Solution Approach 1:
The patent extracts and removes noise and blur components from the captured images through dedicated signal processing operations. By separating and eliminating these detrimental elements while preserving the actual text content, the system maintains ease of capture while significantly improving recognition performance.
Data Source
AI summary
The current document is directed to methods and systems that straighten curvature in the text lines of text-containing digital images, including text-containing digital images generated from the two pages of an open book. Initial processing of a text-containing image identifies the outline of a text-containing page. Next, contours are generated to represent each text line. The midpoints and inclination angles of the links or vectors that comprise the contour lines are determined. A model is constructed for the perspective-induced curvature within the text image. In one implementation, the model, essentially an inclination-angle map, allows for assigning local displacements to pixels within the page image which are then used to straighten the text lines in the text image. In another implementation, the model is essentially a pixel-displacement map which is used to straighten the text lines in the text image.