METHOD FOR GENERATING TYPOGRAPHICAL LINE

• US 20080131000A1
• Filed: 06/27/2007
• Published: 06/05/2008
• Est. Priority Date: 12/01/2006
• Status: Active Grant

• Current Assignee: Compal Electronics Incorporated
• Original Assignee: Compal Electronics Incorporated
• Inventors: Lue, Hsin-Te, Tsai, Wen-Hann
• Granted Patent: US 7,929,772 B2
• Time in Patent Office: Days
• US Class Current: 382/182
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/158 using character size, text spacings or pitch estimation

Abstract

A method for generating typographical line is provided. In the present method, an asymptote of an upper or a lower edge of a line of printing words is obtained first. Then, two typographical lines of the other edge of the line of printing words are obtained according to the asymptote. Two typographical lines of the present edge of the line of printing words are obtained based on the previously obtained typographical lines. Finally, the relations of these typographical lines and edge reference points of the line of printing words are used for removing useless typographical lines. Therefore, the typographical lines obtained by the present invention can provide the means of recognizing word direction, large or small character writing, and punctuation marks, so as to increase the efficiency and accuracy of character recognition.

Claims

1. A method for generating typographical line adapted to generating a plurality of typographical lines of a line of printing words, wherein the line of printing words comprises a plurality of printing characters, the method comprising:

(a) scanning the line of printing words and labeling a first edge and a second edge of each printing character in the line of printing words;

(b) extracting a first edge reference point of the first edge and a second edge reference point of the second edge of each of the printing characters, respectively;

(c) using a least square method to obtain a first straight line asymptotic to the first edge reference points;

(d) using the first straight line as a first base line to calculate a vertical distance between each of the second edge reference points and the first base line;

(e) using a group converging algorithm to divide the second edge reference points into a first group and a second group according to the vertical distances;

(f) using the least square method to obtain a second straight line and a third straight line asymptotic to the first group and the second group of the second edge reference points, respectively;

(g) using the second straight line or the third straight line obtained from corresponding first group or second group that has the most reference points as a second base line to calculate a vertical distance between each of the first edge reference point and the second base line;

(h) using the group converging algorithm to divide the first edge reference points into a third group and a fourth group according to the vertical distances;

(i) using the least square method to obtain a fourth straight line and a fifth straight line asymptotic to the first edge reference points of the third group and the fourth group, respectively; and

(j) using the second straight line, the third straight line, the fourth straight line and the fifth straight line as the typographical lines of the printing word line.

2. The method for generating typographical line according to claim 1, wherein, after the step (a), further comprising:

(a1) recognizing and adjusting a main direction of the line of printing words.

3. The method for generating typographical line according to claim 2, wherein the main direction comprises either a horizontal direction or a vertical direction.

4. The method of generating typographical line according to claim 1, wherein, after the step (a), further comprising:

(a2) calculating a character height of each of the printing characters;

(a3) comparing the character height of each of the printing characters with a preset height and classifying those printing characters having a character height smaller than the preset height as smaller characters.

5. The method of generating typographical line according to claim 4, wherein, after the step (a3), further comprising:

(a4) extracting a central reference point from the center of each of the remaining printing characters;

(a5) using the least square method to obtain a center line asymptotic to the central reference points;

(a6) determining whether a lower edge of each of the printing characters is located above the center line and classifying the printing characters having the lower edge above the center line as small characters; and

(a7) determining whether an upper edge of each of the printing characters is located below the center line and classifying the printing characters having the upper edge below the center line as small characters.

6. The method of generating typographical line according to claim 5, wherein the small characters are not listed as references for generating the typographical lines.

7. The method of generating typographical line according to claim 1, wherein the step (f) comprises:

determining whether the number of second edge reference points in the first group or the second group is 1;

if the number of the second edge reference points in the first group is 1, then the third straight line asymptotic to the second edge reference points of the second group is parallel-shifted to the second reference point of the first group to serve as the second straight line; and

if the number of the second edge reference points in the second group is 1, then the second straight line asymptotic to the second edge reference points of the first group is parallel-shifted to the second reference point of the second group to serve as the third straight line.

8. The method of generating typographical line according to claim 1, wherein the step (i) comprises:

determining whether the number of first edge reference points in the third group or the fourth group is 1;

if the number of the first edge reference points in the third group is 1, then the fifth straight line asymptotic to the first edge reference points of the fourth group is parallel-shifted to the first reference point of the third group to serve as the fourth straight line; and

if the number of the first edge reference points in the fourth group is 1, then the fourth straight line asymptotic to the first edge reference points of the third group is parallel-shifted to the first reference point of the fourth group to serve as the fifth straight line.

9. The method of generating typographical line according to claim 1, wherein a center line of the line of printing words is used as a base, the second straight line is located outside the third straight line and the fifth straight line is located outside the fourth straight line.

10. The method of generating typographical line according to claim 9, wherein, after the step (i), further comprising:

(i1) calculating a first zone distance between the second straight line and the third straight line, a second zone distance between the third straight line and the fourth straight line, and a third zone distance between the fourth straight line and the fifth straight line;

(i2) calculating the number of first reference points, the number of second reference points, the number of third reference points and the number of fourth reference points in the first group, the second group, the third group and the fourth group, respectively; and

(i3) comparing the aforesaid zone distances and the number of reference points so as to remove the aforesaid straight lines.

11. The method of generating typographical line according to claim 10, wherein the step (i3) comprises:

if the first zone distance is smaller than a base value and the number of first reference points is greater than the number of second reference points, then the third straight line is removed.

12. The method of generating typographical line according to claim 10, wherein the step (i3) comprises:

if the third zone distance is smaller than a base value and the number of fourth reference point is greater than the number of third reference points, then the fourth straight line is removed.

13. The method of generating typographical line according to claim 1, wherein, after the step (j), further comprising:

(k) calculating a number of first connected components between the second straight line and the third straight line and a number of second connected components between the fourth straight line and the fifth straight line;

(l) if the number of first connected components is greater than the number of second connected components, then the line of printing words is determined to be erect; and

(m) if the number of first connected components is smaller than the number of second connected components, then the line of printing words is determined to be inverted.

14. The method of generating typographical line according to claim 1, wherein the first edge and the second edge are an upper edge and a lower edge of the printing characters or the opposite.

15. The method of generating typographical line according to claim 1, wherein the first edge reference point and the second edge reference point of each of the printing characters comprise either the center point or end point of the first edge and the second edge of each of the printing characters.

16. The method of generating typographical line according to claim 1, wherein the group converging algorithm comprises a K-Mean algorithm.