Method and System For Character Recognition

• US 20080137971A1
• Filed: 04/01/2005
• Published: 06/12/2008
• Est. Priority Date: 04/01/2004
• Status: Active Grant

• Current Assignee: Kyocera Corporation
• Original Assignee: Exbiblio BV
• Inventors: Kushler, Clifford A., Stafford-Fraser, James Q., King, Martin T., Grover, Dale L.
• Granted Patent: US 9,008,447 B2
• Time in Patent Office: Days
• US Class Current: 382/229
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/186 by deriving mathematical or geometrical properties from the whole image

Abstract

A method and system for character recognition are described. In one embodiment, it may use matched sequences rather than character shape to determine a computer legible result.

Claims

1. ) A method of character recognition comprising:

capture a scan;

determine a matched sequence of unrecognized images; and

determine a computer legible result.

2. ) The method of claim 1 not comprising using language statistics.

3. ) The method of claim 1 not comprising storing information about distances between unrecognized images.

4. ) The method of claim 1 further comprising querying a database.

5. ) The method of claim 1 not comprising logic with character statistics.

6. ) The method of claim 1 not comprising logic with language statistics.

7. ) The method of claim 1 wherein a computer legible result for a given document is usually error free.

8. ) The method of claim 1 wherein a computer legible result for a given document comprises mostly words.

9. ) The method of claim 1 wherein a computer legible result for at least 95% of documents has an error rate smaller than current OCR techniques.

10. ) The method of claim 1 wherein a computer legible result for at least 95% of documents has an error rate smaller than 10%.

11. ) The method of claim 1 wherein a computer legible result for at least 95% of documents is more accurate than a probabilistic method.

12. ) The method of claim 1 not comprising storing an image of an entire document.

13. ) The method of claim 1 wherein a scan comprises an image of less than a full document.

14. ) The method of claim 1 wherein a scan is of less than 20 words.

15. ) The method of claim 1 wherein a computer legible result does not comprise any unknown characters.

16. ) The method of claim 1 wherein a computer legible result does not comprise any probabilistic information.

17. ) The method of claim 1 wherein a computer legible result consists of characters or character representations.

18. ) The method of claim 1 not comprising a substitution cipher.

19. ) The method of claim 1 wherein a matched sequence does not consist of integers.

20. ) The method of claim 1 wherein at least one unrecognized image does not consist of one character.

21. ) The method of claim 1 wherein at least one unrecognized image comprises at least two characters.

22. ) The method of claim 1 further comprising accessing storage comprising at least 10 sequences of at least 4 characters.

23. ) The method of claim 1 wherein the matched sequence comprises 5 identifiers.

24. ) The method of claim 1 wherein the matched sequence comprises offsets.

35. ) The system of claim 3 wherein the document is printed.

25. ) A system comprising:

storage comprising a matched sequence of unrecognized images; and

logic to associate a matched group identifier with a character sequence.

26. ) The system of claim 25 further comprising a power source.

27. ) The system of claim 25 further comprising an input comprising an optical sensor.

28. ) The system of claim 25 wherein at least one character sequence does not consist of one character.

29. ) The system of claim 25 wherein storage further comprises possible results.

30. ) The system of claim 25 wherein storage further comprises at least 10 sequences of at least 4 characters.

31. ) The system of claim 25 wherein the logic does not perform a substitution cipher.

32. ) The system of claim 25 wherein the storage does not comprise single character frequency information.

33. ) The system of claim 25 wherein the storage does not comprise two or three character sequence frequency information.

34. ) A system comprising:

a means for inputting a document;

a means for determining a matched sequence of unrecognized images; and

a means for determining a computer legible result.