Character recognition system

• US 5,307,424 A
• Filed: 12/12/1990
• Issued: 04/26/1994
• Est. Priority Date: 12/12/1990
• Status: Expired due to Term

• Current Assignee: Eberhard Kuehl
• Original Assignee: Eberhard Kuehl
• Inventors: Kuehl, Eberhard
• Primary Examiner(s): Moore, David K.
• Assistant Examiner(s): Cammarata, Michael
• Application Number:
  US07/626,517
• Time in Patent Office: 1,231 Days
• Field of Search: 382/10, 382/25, 382/26, 382/21, 382/22, 382/49, 382/27, 382/56, 382/55
• US Class Current: 382/198
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/1444 Selective acquisition, locating or processing of specific regions, e.g. highlighted text, fiducial marks or predetermined fields
• CPC Class Codes: G06V 30/184 by analysing segments intersecting the pattern
• CPC Class Codes: G06V 30/196 using sequential comparisons of the image signals with a plurality of references

Abstract

Characters to be recognized and which may be handwritten are each described by a pattern of one bit thresholded pixel data which follows the center line of the character and encoded into strings of digital data signals (bytes) which represent the shape of the pattern from three aspects, namely the top view, the front edge and the rear edge thereof, thereby providing low level recognition information in the form of three digital data strings. From these strings, the characters are recognized by finding high level recognition strings of digital signals which correspond to different characters in the character set to be recognized. This high level recognition is carried on in two steps, namely allocating the high level recognition strings which correspond to the low level strings for each aspect of the character and then finding combinations of high level strings reflecting the same character, thereby accounting for variations in the shapes in which the same character can be written. The high level strings and the combinations are stored in tables to enable recognition of characters in a character set.

Claims

1. The method of character recognition of characters in a set of characters in which a pattern of pixels in adjacent rows corresponding to the character is provided, which method comprises the steps of encoding the pattern of pixels into a plurality of first strings of digital signals having values dependent upon the shape of a pattern in accordance with the movement of edges of the pattern measured by the relative displacement of the pixels in the adjacent rows to each other to represent the spatial relationship including the connectivity and direction of neighboring ones of said pixels as viewed from a plurality of aspects, translating said first strings into a plurality of second strings of digital signals each representing a different aspect of the different characters in said set such that variations in the shape of the pattern of the same character are included in said second strings, and recognizing the pattern as representing a character in the set when said second strings constitute a combination of second strings corresponding to that character.

2. The method according to claim 1 wherein said recognizing step is carried out by relating said pluralities of second strings to pluralities of third strings which correspond to characters in said set, and recognizing a character in said set when said second strings are related to a plurality of said third strings which correspond to that character.

3. The method according to claim 2 further comprising the step of obtaining said patterns by scanning a page containing said characters to generate said pixels in rows and columns of a matrix in rectangular coordinates, allocating said pixels to different addresses in a memory, selecting different addresses containing pixels for each character to be recognized thereby including each character to be recognized in a frame, and adjusting the coordinates of said frame to compensate for skew of the page while it is being scanned.

4. The method according to claim 1 wherein prior to said encoding step, the step of arranging said pattern and memory in a matrix having rectangular (x and y coordinates) is first carried out.

5. The method according to claim 4 including the step of thresholding said pixels to represent each pixel which has a value greater than a certain value as a single bit representing the presence of a pixel in said pattern.

6. The method according to claim 4 further comprising the step of thinning said pattern in said memory so that it contains data representing pixels along the center line of said pattern.

7. The method according to claim 6 wherein said matrix contains rows and columns along the x and y coordinates thereof, respectively, and said aspects are views of the top, right and left sides of said pattern, said encoding step being carried out by encoding the top view into a string of data signals for each row representing short lines, long lines and holes each by a data signal of different value, and by encoding the left and right sides in accordance with the displacement of pixels in the direction along the columns into strings of data signals of different value in accordance with the amount of such displacement.

8. The method according to claim 7 wherein said displacements for every pixel row for the right and left sides are calculated by computing the displacement delta X in accordance with the equation

9. The method according to claim 1 further comprising the step of storing first, second and third tables containing pluralities of said second strings of digital signals, each plurality corresponding to a different character in said set for a different one of said aspects, and said translating step comprises selecting from said tables said second strings corresponding to said first strings.

10. The method according to claim 9 further comprising the step of storing first, second and third tables containing pluralities of said second strings, each corresponding to different ones of said characters in said set for said top, right and left sides, and said translating step comprises selecting different ones of said second strings which correspond to said first strings for said top, right and left sides, respectively.

11. The method according to claim 9 further comprising the step of storing a fourth table containing third strings corresponding to different combinations of said second strings, said third string combinations corresponding to different characters in said set, and said recognizing step is carried out by selecting said combinations of third strings which correspond to said combinations of second strings, and indicating the character of the select combination of third strings as the character depicted by said pattern.

12. The method according to claim 10 further comprising the step of storing a fourth table containing third strings corresponding to different combinations of said second strings, said third string combinations corresponding to different characters in said set, and said recognizing step is carried out by selecting said combinations of third strings which correspond to said combinations of second strings, and indicating the character of the select combination of third strings as the character depicted by said pattern.

27. The method of claim 1 wherein said translating step is carried out to provide said digital signals in said second strings with fewer bytes than the digital signals of said first strings.

13. Apparatus for optical character recognition of characters in a set of characters in which a pattern of pixels in adjacent rows corresponding to the character is provided, which comprises means for encoding the pattern of pixels into a plurality of first strings of digital signals having values dependent upon the shape of the pattern in accordance with the movement of edges of the pattern measured by the displacement of the pixels in the adjacent rows to each other to represent the spatial relationship including the connectivity and direction of neighboring ones of said pixels as viewed from a plurality of aspects, means for translating said first strings into a plurality of second strings of digital signals each representing a different aspect of the different characters in said set such that variations in the shape of the pattern of the same character are included in said second strings, and means for recognizing the pattern as representing a character in the set when said second strings constitute a combination of second strings corresponding to that character.

14. The apparatus according to claim 13 wherein said recognizing means comprises means for relating said pluralities of second strings to pluralities of third strings which correspond to characters in said set, and means for recognizing a character in said set when said second strings are related to a plurality of said third strings which correspond to that character.

15. The apparatus according to claim 14 further comprising means for scanning a page containing said characters to generate said pixels in rows and columns of a matrix in rectangular coordinates, means for allocating said pixels to different addresses in a memory, means for selecting different addresses containing pixels for each character to be recognized thereby including each character to be recognized in a frame, and means for adjusting the coordinates of said frame to compensate for skew of the page while it is being scanned.

16. The apparatus according to claim 13 further comprising computer memory means, and means for arranging said pattern in said memory in a matrix having rectangular (x and y coordinates) is first carried out.

17. The apparatus according to claim 16 including means for thresholding said pixel values to represent each pixel which has a value greater than a certain value as a single bit representing the presence of a pixel i said pattern.

18. The apparatus according to claim 16 further comprising means for thinning said pattern in said memory so that is contains data representing pixels along the center line of said pattern.

19. The apparatus according to claim 18 wherein said matrix contains rows and columns along the x and y coordinates thereof, respectively, and said aspects are views of the top, right and left sides of said pattern, said encoding means comprising means for encoding the top view into a string of data signals for each row representing short liens, long lines and holes each by a data signal of different value, and for encoding the left and right sides in accordance with the displacement of pixels in the direction along the columns into strings of data signals of different value in accordance with the amount of such displacement.

20. The apparatus according to claim 19 wherein said encoding means includes means for calculating the displacements for every pixel row for the right and left sides which the displacement delta X is determined in accordance with the equation

21. The apparatus according to claim 13 further comprising means for storing first, second and third tables containing pluralities of said second strings of digital signals, each plurality corresponding to a different character in said set for a different one of said aspects, and said translating means comprises means for selecting from said tables said second strings corresponding to said first strings.

22. The apparatus according to claim 21 further comprising means for storing first, second and third tables containing pluralities of said second strings, each corresponding to different ones of said characters in said set for said top, right and left sides, and said translating means comprises means for selecting different ones of said second strings which correspond to said first strings for said top, right and left sides, respectively.

23. The apparatus according to claim 21 further comprising means for storing a fourth table containing third strings corresponding to different combinations of said second strings, said third string combinations corresponding to different characters in said set, and said recognizing means comprises means for selecting said combinations of third strings which correspond to said combinations of second strings, and means for indicating the character of the select combination of third strings as the character depicted by said pattern.

24. The apparatus according to claim 22 further comprising means for storing a fourth table containing third strings corresponding to different combinations of said second strings, said third string combinations corresponding to different characters in said set, and said recognizing means comprises means for selecting said combinations of third strings which correspond to said combinations of second strings, and means for indicating the character of the select combination of third strings as the character depicted by said pattern.

28. The apparatus according to claim 13, wherein said translating means includes means for providing said digital signals in said second strings with fewer bytes than said digital signals of said first strings.

25. Apparatus for optical character recognition of handwritten characters in a set thereof which comprises means for optically generating and electronically storing a pattern of pixel data items in adjacent rows in two dimensions representing a character to be recognized, means for encoding the pattern into a first group of data bytes representing the pattern in accordance with the movement of edges of the pattern measured by the relative displacement of the pixels in the adjacent rows to each other to represent the geometric relationship including the connectivity and direction of neighboring pixels thereof and where the bytes in the first group are a reduction in number of the data items in the pattern, means for translating and converting different bytes in said first group which are common to certain character shapes into other bytes corresponding to the characters in said set and fewer in number than the number of byes in said first group, and means for recognizing a pattern as representing a character in said set when a plurality of the bytes in said second group constitute, in the order in which they occur in said second group, a combination corresponding to that character.

26. The method for the recognition of handwritten characters in a set thereof which comprises the steps of optically generating and electronically storing a pattern of pixel data items in adjacent rows in two dimensions representing a character to be recognized, then encoding the pattern into a first group of data bytes representing the pattern in accordance with the movement of edges of the pattern measured by the relative displacement of the pixels in the adjacent rows to each other to represent the geometric relationship including the connectivity and direction of neighboring pixels thereof and where the bytes in the first group are a reduction in the number of data items in the pattern, then translating and converting different bytes to said first group which are common to certain shapes common to different characters in said set into other bytes representing said shapes to form a second group of data bytes corresponding to characters in said set and fewer in number than the bytes in said first group, and recognizing the pattern as representing a character in said set when a plurality of the bytes in said second group constitute in the order in which they occur in said second group, a combination corresponding to that character.