METHOD FOR RECOGNIZING CHARACTERS

• US 20080137955A1
• Filed: 06/27/2007
• Published: 06/12/2008
• Est. Priority Date: 12/06/2006
• Status: Active Grant

• Current Assignee: Compal Electronics Incorporated
• Original Assignee: Compal Electronics Incorporated
• Inventors: Tsai, Wen-Hann, Fann, Sheng-Ann
• Granted Patent: US 7,929,773 B2
• Time in Patent Office: Days
• US Class Current: 382/176
• CPC Class Codes: G06V 30/244 using graphical properties, e.g. alphabet type or font

Abstract

A method for recognizing characters is provided. All the characters to be recognized are categorized into a plurality of character categories according to the relative positions of the characters in the typographical lines before the characters are recognized. When recognizing the characters, a character is compared with only those characters in the corresponding character category of the character in the character database. Therefore, the range and number of characters to be compared are reduced, and the accuracy and speed for recognizing characters are improved.

Claims

1. A character recognition method, comprising:

a. scanning a line of printing words, wherein the line of printing words comprises a plurality of first characters;

b. generating a plurality of typographical lines by using the first characters;

c. determining a character category of each first character according to a relative position of the first character in the typographical lines; and

d. comparing each first character with a plurality of second characters in a database belonging to the corresponding character category of the first character to obtain the second character corresponding to the first character and accordingly to recognize the first character, wherein the database records a plurality of character categories and the second characters in each of the character categories.

2. The character recognition method according to claim 1, wherein the typographical lines comprise a top line, an upper line, a base line, and a bottom line, the area between the top line and the upper line is an upper zone, the area between the upper line and the base line is a central zone, and the area between the base line and the bottom line is a downer zone.

3. The character recognition method according to claim 2, wherein step c. comprises:

c1. determining whether or not each first character is a small character;

c2. performing a small character categorization if the first character is a small character; and

c3. performing a non-small character categorization if the first character is not a small character.

4. The character recognition method according to claim 3, wherein step c1. comprises:

c1-1. calculating a character height of each first character; and

c1-2. comparing the character height of each first character with a predetermined height, and categorizing the first character as a small character if the character height of the first character is smaller than the predetermined height.

5. The character recognition method according to claim 4, wherein after step c1-2., the character recognition method further comprises:

c1-3. respectively capturing a central reference point of each of the remaining first characters;

c1-4. obtaining a central line asymptotic to the central reference points through a least square method;

c1-5. determining whether or not a lower edge of each first character is located above the central line, and categorizing the first character as a small character if the lower edge of the first character is located above the central line; and

c1-6. determining whether or not an upper edge of each first character is located below the central line, and categorizing the first character as a small character if the upper edge of the first character is located below the central line.

6. The character recognition method according to claim 3, wherein step c. further comprises:

categorizing the typographical lines into one of a plurality of states according to the number and types of the typographical lines.

7. The character recognition method according to claim 6, wherein the states comprise:

a first state, denoting that the typographical lines comprise the top line, the upper line, the base line, and the bottom line;

a second state, denoting that the typographical lines comprise the base line, the bottom line, and the typographical line combined by the top line and the upper line;

a third state, denoting that the typographical lines comprise the top line, the upper line, and the typographical line combined by the base line and the bottom line; and

a fourth state, denoting that the typographical lines comprise the typographical line combined by the top line and the upper line and the typographical line combined by the base line and the bottom line.

8. The character recognition method according to claim 7, wherein the categorization of each small character comprises:

c2-1. locating a central line between the top line and the upper line, and calculating a first distance, a second distance, a third distance, a fourth distance, and a fifth distance from a central reference point of the small character to the top line, the upper line, the base line, and the central line;

c2-2. calculating a ratio between a height and a width of the small character;

c2-3. determining whether or not the ratio is greater than a first threshold;

c2-4. determining whether or not the typographical lines are categorized to the first state or the second state if the ratio is greater than the first threshold;

c2-4-1. calculating distances from the lower edge of the small character to the top line and the upper line and determining whether or not the lower edge of the small character is closer to the upper line than to the top line if the typographical lines are categorized to the first state or the second state;

c2-4-1-1. categorizing the small character to a fourth category if the lower edge of the small character is closer to the upper line than to the top line;

c2-4-1-2. categorizing the small character to a second category if the lower edge of the small character is not closer to the upper line than to the top line;

c2-4-2. categorizing the small character to an eighth category if the typographical lines are not categorized to the first state and the second state;

c2-5. determining whether or not the fifth distance is smaller than a second threshold or whether or not the fifth distance is smaller than the second distance and the third distance if the ratio is not greater than the first threshold;

c2-5-1. determining whether or not the third distance is smaller than the fifth distance and the second distance if the fifth distance is smaller than a second threshold or the fifth distance is smaller than the second distance and the third distance;

c2-5-1-1. categorizing the small character to a sixth category if the third distance is smaller than the fifth distance and the second distance;

c2-5-1-2. determining whether or not the second distance is smaller than the fifth distance and the third distance if the third distance is not smaller than the fifth distance or the second distance;

c2-5-1-2-1. categorizing the small character to a fifth category if the second distance is smaller than the fifth distance and the third distance;

c2-5-1-2-2. categorizing the small character to a seventh category if the second distance is not smaller than the fifth distance or the third distance;

c2-5-2. determining whether or not the central reference point is located on the central line if the fifth distance is not smaller than a second threshold and the fifth distance is not smaller than the second distance or the third distance;

c2-5-2-1. calculating distances from the lower edge of the small character to the upper line and the base line and determining whether or not the lower edge of the small character is closer to the upper line than to the base line if the central reference point is located on the central line;

c2-5-2-1-1. categorizing the small character to the fifth category if the lower edge of the small character is closer to the upper line than to the base line;

c2-5-2-1-2. categorizing the small character to the fourth category if the lower edge of the small character is not closer to the upper line than to the base line; and

c2-5-2-2. categorizing the small character to the sixth category if the central reference point is not located on the central line.

9. The character recognition method according to claim 8, wherein the categorization of the non-small character comprises:

c3-1. determining whether or not the upper edge of the first character is located above the top line if the typographical lines are categorized to the first state;

c3-1-1. determining whether or not the lower edge of the first character is located above the bottom line if the upper edge of the first character is located above the top line;

c3-1-1-1. categorizing the first character to a first category if the lower edge of the first character is located above the bottom line;

c3-1-1-2. categorizing the first character to a second category if the lower edge of the first character is not located above the bottom line;

c3-1-2. determining whether or not the lower edge of the first character is located above the bottom line if the upper edge of the first character is not located above the top line;

c3-1-2-1. categorizing the first character to a third category if the lower edge of the first character is located above the bottom line; and

c3-1-2-2. categorizing the first character to the fourth category if the lower edge of the first character is not located above the bottom line.

10. The character recognition method according to claim 9, wherein the categorization of the non-small character further comprises:

c3-2. determining whether or not the upper edge of the first character is located above the top line if the typographical lines are categorized to the second state;

c3-2-1. categorizing the first character to the second category if the upper edge of the first character is located above the top line; and

c3-2-2. categorizing the first character to the fourth category if the upper edge of the first character is not located above the top line.

11. The character recognition method according to claim 10, wherein the categorization of the non-small character further comprises:

c3-3. categorizing the first character to the eighth category if the typographical lines are categorized to the third state or the fourth state.

12. The character recognition method according to claim 11, wherein the character categories comprise:

the first category taking the upper zone, the central zone, and the downer zone;

the second category taking the upper zone and the central zone;

the third category taking the downer zone;

the fourth category taking the central zone;

the fifth category being the small character, the fifth category being located around the central line;

the sixth category taking the area between the central line and the upper line;

the seventh category taking the area between the central line and the base line; and

the eighth category comprising the remaining characters cannot be categorized according to the typographical lines.

13. The character recognition method according to claim 1, wherein step d. comprises:

d1. calculating a first characteristic value of each first character; and

d2. comparing the first characteristic value of each first character with a second characteristic value of each second character in the corresponding character category of the first character in the database to obtain the second character having the closest characteristic value as a recognized character of the first character.

14. The character recognition method according to claim 13, wherein step d2. further comprises:

recognizing each first character by using a corresponding prediction model of the first character.