METHOD OF COMPENSATING FOR DISTORTION IN TEXT RECOGNITION

• US 20100208999A1
• Filed: 02/12/2010
• Published: 08/19/2010
• Est. Priority Date: 02/13/2009
• Status: Active Grant

• Current Assignee: Industry Foundation Of Chonnam National University (Chonnam National University), Samsung Electronics Co. Ltd.
• Original Assignee: Industry Foundation Of Chonnam National University (Chonnam National University), Samsung Electronics Co. Ltd.
• Inventors: Kim, Soo-Hyung, Lee, Guee-Sang, Yang, Hyung-Jeong, Kim, Eui-Chul, Hwang, Seong-Taek, Oh, Sang-Wook, Kim, Hyun-Soo, Kim, Sang-Ho
• Granted Patent: US 8,417,057 B2
• Time in Patent Office: Days
• US Class Current: 382/195
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/146 Aligning or centring of the image pick-up or image-field
• CPC Class Codes: G06V 30/15 Cutting or merging image elements, e.g. region growing, watershed or clustering-based techniques
• CPC Class Codes: G06V 30/1607 Correcting image deformation, e.g. trapezoidal deformation caused by perspective

Abstract

A method of compensating for distortion in text recognition is provided, which includes extracting a text region from an image; estimating the form of an upper end of the extracted text region; estimating the form of a lower end of the extracted text region; estimating the form of left and right sides of the extracted text region; estimating a diagram constituted in the form of the estimated upper end, lower end, left and right sides, and including a minimum area of the text region; and transforming the text region constituting the estimated diagram into a rectangular diagram using an affine transform.

Claims

1. A method of compensating for distortion in text recognition, the method comprising the steps of:

extracting a text region from an image;

estimating a form of an upper end of the extracted text region;

estimating a form of a lower end of the extracted text region;

estimating a form of left and right sides of the extracted text region;

estimating a diagram in the form of the estimated upper end, lower end, and left and right sides of the extracted text region, including a minimum area of the text region; and

transforming the text region constituting the estimated diagram into a rectangular diagram using an affine transform.

2. The method of claim 1, wherein the step of estimating the form of the upper end comprises:

positioning a first reference line that simultaneously passes through individual text regions constituting an entire text region, and extracting coordinate values of pixels corresponding to an uppermost end among pixels positioned at upper ends that are higher than the first reference line; and

obtaining a straight line approximating the extracted coordinate values, and estimating the obtained straight line in the form of the upper end of the extracted text region.

3. The method of claim 2, wherein the approximating straight line corresponds to a straight line in the following Equation (1) that is obtained through the following Equations (2) and (3), where “

x” and

“

y”

denote coordinate values of the respective pixels, and “

n”

denotes a number of uppermost pixels positioned at upper ends that are higher than the first reference line

$\begin{array}{cc}y=\mathrm{mx}+b& \left(1\right)\\ m=\frac{n<br><br>\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>x_i<br><br>y_i-\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>x_i<br><br>\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>y_i}{n<br><br>\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>x_i^2-{\left(\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>x_i\right)}^2}& \left(2\right)\\ b=\frac{\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>x_i^2<br><br>\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>y_i-n<br><br>\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>x_i<br><br>y_i<br><br>\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>x_i}{n<br><br>\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>x_i^2-{\left(\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>x_i\right)}^2}.& \left(3\right)\end{array}$

4. The method of claim 1, wherein the step of estimating the form of the lower end comprises the steps of:

positioning a second reference line that simultaneously passes through individual text regions constituting an entire text region, and extracting coordinate values of pixels corresponding to the lowermost end among pixels positioned at lower ends that are higher than the second reference line; and

obtaining a straight line approximating the extracted coordinate values, and estimating the obtained straight line in the form of the lower end of the extracted text region.

5. The method of claim 4, wherein the approximating straight line corresponds to a straight line in the following Equation (1) that is obtained through the following Equations (2) and (3), where “

x” and

“

y”

denote coordinate values of the respective pixels, and “

n”

denotes the number of uppermost pixels positioned at upper ends that are higher than the first reference line

$\begin{array}{cc}y=\mathrm{mx}+b& \left(1\right)\\ m=\frac{n<br><br>\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>x_i<br><br>y_i-\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>x_i<br><br>\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>y_i}{n<br><br>\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>x_i^2-{\left(\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>x_i\right)}^2}& \left(2\right)\\ b=\frac{\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>x_i^2<br><br>\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>y_i-n<br><br>\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>x_i<br><br>y_i<br><br>\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>x_i}{n<br><br>\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>x_i^2-{\left(\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>x_i\right)}^2}.& \left(3\right)\end{array}$

6. The method of claim 1, wherein the step of estimating the form of the left and right sides comprises:

measuring slopes of contours of individual text regions positioned on the left and right sides in an entire text region; and

estimating slopes of contours of the individual text regions, which correspond to the slopes having a greatest distribution as the result of measurement, in the form of the left and right sides.

7. The method of claim 1, wherein the step of estimating the form of the left and right sides comprises:

measuring slopes of contours of individual text regions positioned on the left and right sides in an entire text region; and

estimating the slopes of the individual text regions having a largest size in the form of the left and right sides.

8. The method of claim 6, wherein the slopes of the individual text regions estimated in the form of the left and right sides are in a range excluding a range of −

2°

to +2°

on the basis of a vertical line.

9. The method of claim 7, wherein the slopes of the individual text regions estimated in the form of the left and right sides are in a range excluding a range of −

2°

to +2°

on the basis of a vertical line.