Positioning templates in optical character recognition systems

• US 6,081,621 A
• Filed: 10/01/1996
• Issued: 06/27/2000
• Est. Priority Date: 10/01/1996
• Status: Expired due to Term

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Ackner, Reuven
• Primary Examiner(s): Cuoso, Yon J.
• Application Number:
  US08/724,478
• Time in Patent Office: 1,365 Days
• Field of Search: 382/161, 382/206, 382/209, 382/216, 382/218, 382/229, 382/278, 382/295, 382/160, 382/226, 382/230
• US Class Current: 382/216
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/148 Segmentation of character regions
• CPC Class Codes: G06V 30/1902 Shifting or otherwise transforming the patterns to accommodate for positional errors
• CPC Class Codes: G06V 30/2253 Recognition of characters printed with magnetic ink G06V30/2247 takes precedence

Abstract

One dimensional horizontal and vertical reference projections are obtained for a reference character template. Similar one dimensional horizontal and vertical pattern projections are obtained for an unknown character pattern. The horizontal reference projection and horizontal pattern projection are compared, and the vertical reference projection and vertical pattern projection are compared. The reference character template is then positioned with respect to the unknown character pattern based on such comparisons in order to calculate a correlation coefficient between the two. In certain cases, the unknown character can be identified based solely on its one dimensional projections. Multiple characters in a character pattern are identified either one at a time or as a group. In the latter case, all possible combinations are generated using tree search processing, and the best combination is selected.

Claims

1. A method for identifying plural characters in an unknown character pattern, the unknown character pattern comprising two-dimensional pixel data, said method comprising:

a first generating step of generating potentially matching characters for a first character position in the unknown character pattern;

a second generating step of generating potentially matching characters for a subsequent character position, wherein a spatial location of the subsequent character position is determined in accordance with a potentially matching character string for all previous character positions in the unknown character pattern, and wherein each potentially matching character generated for the subsequent character position is linked to the potentially matching character string, thereby forming plural new potentially matching character strings; and

an identifying step of identifying the characters in the unknown character pattern in accordance with one of the potentially matching character strings that is most similar to the unknown character pattern.

2. A method according to claim 1, wherein when a portion of the unknown character pattern has been previously explored and a tentative identification for said portion of the unknown character pattern is again required, a corresponding previously identified structure is copied to a current character position.

3. A method according to claim 1, wherein potentially matching characters are generated by aligning each of plural reference character templates with the unknown character pattern at a current character position, calculating a correlation coefficient between each such reference character template and said unknown character pattern, as so aligned, and selecting those reference character templates having correlation coefficients higher than a specified threshold.

4. A method according to claim 1, further comprising calculating a correlation coefficient between a template corresponding to each potentially matching character and the unknown character pattern, wherein the characters in the unknown character pattern are identified in accordance with said correlation coefficients.

5. An apparatus for identifying plural characters in an unknown character pattern, the unknown character pattern comprising two-dimensional pixel data, said apparatus comprising:

a memory for storing the plural characters and program instruction sequences; and

a processor for executing said program instruction sequences stored in said memory,

wherein said program instruction sequences include (a) a first generating step to generate potentially matching characters for a first character position in the unknown character pattern, (b) a second generating step to generate potentially matching characters for a subsequent character position, wherein a spatial location of the subsequent character position is determined in accordance with a potentially matching character string for all previous character positions in the unknown character pattern, and wherein each potentially matching character generated for the subsequent character position is linked to the potentially matching character string, thereby forming plural new potentially matching character strings, and (c) an identifying step to identify the characters in the unknown character pattern in accordance with one of the potentially matching character strings that is most similar to the unknown character pattern.

6. An apparatus according to claim 5, wherein said program instruction sequences further include a step to copy a previously identified structure to a current character position.

7. An apparatus according to claim 5, wherein each of the first and second generating steps includes a step to align each of plural reference character templates with the unknown character pattern at a current character position, a step to calculate a correlation coefficient between each such reference character template and said unknown character pattern, as so aligned, and a step to select those reference character templates having correlation coefficients higher than a specified threshold.

8. An apparatus according to claim 5, wherein said program instruction sequences further include a step to calculate a correlation coefficient between a template corresponding to each potentially matching character and the unknown character pattern, wherein the characters in the unknown character pattern are identified in accordance with said correlation coefficients.

9. Computer executable process steps stored on a computer readable medium to identify plural characters in an unknown character pattern, the unknown character pattern comprising two-dimensional pixel data, the process steps comprising:

a first generating step to generate potentially matching characters for a first character position in the unknown character pattern;

a second generating step to generate potentially matching characters for a subsequent character position, wherein a spatial location of the subsequent character position is determined in accordance with a potentially matching character string for all previous character positions in the unknown character pattern, and wherein each potentially matching character generated for the subsequent character position is linked to the potentially matching character string, thereby forming plural new potentially matching character strings; and

an identifying step to identify the characters in the unknown character pattern in accordance with one of the potentially matching character strings that is most similar to the unknown character pattern.

10. Computer executable process steps according to claim 9, further comprising a step to copy a previously identified structure to a current character position.

11. Computer executable process steps according to claim 9, wherein each of the first and second generating steps includes a step to align each of plural reference character templates with the unknown character pattern at a current character position, a step to calculate a correlation coefficient between each such reference character template and said unknown character pattern, as so aligned, and a step to select those reference character templates having correlation coefficients higher than a specified threshold.

12. Computer executable process steps according to claim 9, further comprising a step to calculate a correlation coefficient between a template corresponding to each potentially matching character and the unknown character pattern, wherein the characters in the unknown character pattern are identified in accordance with said correlation coefficients.