Image processing method and an image processing apparatus

• US 5,835,632 A
• Filed: 03/07/1996
• Issued: 11/10/1998
• Est. Priority Date: 03/08/1995
• Status: Expired due to Fees

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Yoshii, Hiroto, Arai, Tsunekazu, Takasu, Eiji
• Primary Examiner(s): Couso, Jose L.
• Assistant Examiner(s): Bella, Matthew C.
• Application Number:
  US08/612,494
• Time in Patent Office: 978 Days
• Field of Search: 382/159, 382/161, 382/182, 382/185, 382/186, 382/187, 382/188, 382/189, 382/198, 382/200, 382/218
• US Class Current: 382/185
• CPC Class Codes: G06F 18/28 Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/166 Normalisation of pattern dimensions
• CPC Class Codes: G06V 30/1914 Determining representative reference patterns, e.g. averaging or distorting patterns; Generating dictionaries, e.g. user dictionaries

Abstract

An image processing system includes entering a predetermined sample character, comparing characteristic data for the sample character with characteristic data for a predetermined reference character, extracting characteristic data inherent to the entered sample character and relative to the reference character, storing the extracted characteristic data, and correcting a character which is newly entered for recognition and different from the predetermined sample character in accordance with the stored characteristic data.

Claims

1. An image processing apparatus comprising:

input means for inputting a pattern representing a predetermined sample character;

extraction means for comparing the pattern input by said input means with a reference pattern representing the predetermined sample character, and for extracting characteristic data inherent to the input pattern in accordance with a comparison result;

holding means for holding the characteristic data extracted by said extraction means, as a correction rule common to a plurality of different patterns representing respective characters to be recognized; and

correction means for correcting a pattern entered for recognition after said holding means has held the characteristic data and different from the input pattern representing the predetermined sample character, in accordance with the characteristic data held by said holding means.

2. An image processing apparatus according to claim 1, wherein said input means includes display means for displaying sample characters with which to prompt character input by an operator.

3. An image processing apparatus according to claim 1, wherein said extraction means employs characteristic values for said sample characters that are input at said input means so as to extract, as said character shape data, a largest rotation angle for a recognition result value that is obtained when a rotation angle of each of said sample characters is sequentially changed.

4. An image processing apparatus according to claim 3, wherein said correction means rotates said character that is input in consonance with said rotation angle that serves as said characteristic data that are held by said holding means.

5. An image processing apparatus according to claim 1, wherein said extraction means computes a tilt of said sample characters, which are input at said input means, from said characteristic data for said sample characters.

6. An image processing apparatus according to claim 5, wherein said correction means corrects for a tilt of said characteristic data of said character that is input, in consonance with tilt data that serve as said characteristic data that are held by said holding means.

7. An image processing apparatus according to claim 1, wherein said input means inputs a plurality of sample characters and wherein said extraction means further includes means for computing an average shape for said characteristic data for each of said sample characters.

8. An image processing apparatus according to claim 1, wherein said holding means is provided in a nonvolatile storage device.

9. An image processing apparatus according to claim 1, wherein said correction means extracts said characteristic data for characters that are input and performs a correction for said characteristic data that are extracted.

10. An image processing apparatus according to claim 1, wherein said input means is coordinate input means.

11. An image processing apparatus according to claim 1, wherein said input means is a scanner.

12. An image processing method comprising:

an input step of inputting a pattern representing a predetermined sample character;

an extraction step of comparing the pattern input in said input step with a reference pattern representing the predetermined sample character, and of extracting characteristic data inherent to the input pattern in accordance with a comparison result;

a storage step of storing in a predetermined holding means the characteristic data extracted in said extraction step, as a correction rule common to a plurality of different patterns representing respective characters to be recognized; and

a correction step of correcting a pattern entered for recognition after said holding step has held the characteristic data and different from the input pattern representing the predetermined sample character, in accordance with the characteristic data held by said holding means.

13. An image processing method according to claim 12, wherein said input step includes a display step of displaying sample characters with which to prompt character input by an operator.

14. An image processing method according to claim 12, wherein, at said extraction step, characteristic values for said sample characters that are input at said input step are employed so as to extract, as said character shape data, a largest rotation angle for a recognition result value that is obtained when a rotation angle of each of said sample characters is sequentially changed.

15. An image processing method according to claim 14, wherein, at said correction step, said character that is input is rotated in consonance with said rotation angle that serves as said characteristic data that are held by said holding means.

16. An image processing method according to claim 12, wherein, at said extraction step, a tilt of said sample characters, which are input at said input means, is computed from said characteristic data for said sample characters.

17. An image processing method according to claim 16, wherein, at said correction step, a tilt of said characteristic data of said character that is input, is corrected for in consonance with tilt data that serve as said characteristic data that are held by said holding means.

18. An image processing method according to claim 12, wherein a plurality of sample characters are input at said input step, and wherein said extraction step further includes a step of computing an average shape for said characteristic data for each of said sample characters.

19. An image processing method according to claim 12, wherein said holding means is provided in a nonvolatile storage device.

20. An image processing method according to claim 12, wherein, at said correction step, said characteristic data for characters that are input are extracted and a correction is performed for said characteristic data that are extracted.

21. An image processing apparatus comprising:

input means for inputting a pattern representing a predetermined sample character;

extraction means for comparing the pattern input by said input means with a reference pattern representing the predetermined sample character, and for extracting character shape data inherent to an operator in accordance with a comparison result;

holding means for holding the character shape data extracted by said extraction means, as a correction rule common to a plurality of different patterns representing respective characters to be recognized; and

correction means for correcting characteristic data for a handwritten character entered for recognition after said holding means has held the characteristic shape data and different from the input pattern representing the predetermined sample character, in accordance with the character shape data held by said holding means; and

character recognition means for recognizing the characteristic data corrected by said correction means.

22. An image processing apparatus according to claim 21, wherein said input means includes display means for displaying sample characters with which to prompt character input by an operator.

23. An image processing apparatus according to claim 21, wherein said extraction means employs characteristic values for said sample characters that are input at said input means so as to extract, as said character shape data, a largest rotation angle for a recognition result value that is obtained when a rotation angle of each of said sample characters is sequentially changed.

24. An image processing apparatus according to claim 23, wherein said correction means rotates said characteristic data for said character that is input in consonance with said rotation angle that serves as said character shape data that are held by said holding means.

25. An image processing apparatus according to claim 21, wherein said extraction means computes a tilt of said sample character, which is input at said input means, from said characteristic values for said sample character.

26. An image processing apparatus according to claim 25, wherein said correction means corrects data that is related to a tilt of said characteristic data of said character that is input, in consonance with tilt data that serve as said character shape data that are held by said holding means.

27. An image processing apparatus according to claim 21, wherein said input means inputs a plurality of sample characters and wherein said extraction means further includes means for computing an average shape for said character shape data for each of said sample characters.

28. An image processing apparatus according to claim 21, wherein said holding means is provided in a nonvolatile storage device.

29. An image processing apparatus according to claim 21, wherein said input means is coordinate input means.

30. An image processing apparatus according to claim 21, further comprising display means for displaying a character pattern that is obtained as a result of a character recognition process performed by said character recognition means.

31. A computer-readable memory medium which stores computer-executable process steps, the computer-executable process steps to process an image, the computer-executable process steps comprising:

an input step to input a pattern representing a predetermined sample character;

an extraction step to comparing the pattern input in the input step with a reference pattern representing the predetermined sample character, and to extract characteristic data inherent to the input pattern in accordance with a comparison result;

a storage step to store in a predetermined holding means the characteristic data extracted in the extraction step, as a correction rule common to a plurality of different patterns representing respective characters to be recognized; and

a correction step correct a pattern entered for recognition after the holding step has held the characteristic data and different from the input pattern representing the predetermined sample character, in accordance with the characteristic data held by the holding means.

32. A computer-readable memory medium which stores computer-executable process steps, the computer-executable process steps to process an image, the computer-executable process steps comprising:

an input step to input a pattern representing a predetermined sample character;

an extraction step to compare the pattern input by the input means with a reference pattern representing the predetermined sample character, and to extract character shape data inherent to an operator in accordance with a comparison result;

a holding step for holding the character shape data extracted by the extraction means, as a correction rule common to a plurality of different patterns representing respective characters to be recognized; and

a correcting step to correct characteristic data for a handwritten character entered for recognition after the holding means has held the characteristic shape data and different from the input pattern representing the predetermined sample character, in accordance with the character shape data held the holding step; and

a recognizing step to recognize the characteristic data corrected in the correcting step.