Method and apparatus for developing a character recognition dictionary

• US 6,208,755 B1
• Filed: 01/19/1995
• Issued: 03/27/2001
• Est. Priority Date: 01/26/1994
• Status: Expired due to Term

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Harada, Takashi, Matsubayashi, Kazuhiro, Yoshii, Hiroto, Arai, Tsunekazu, Mori, Shigeki, Takasu, Eiji, Sakaguchi, Katsuhiko
• Primary Examiner(s): COUSO, YON JUNG
• Application Number:
  US08/374,819
• Time in Patent Office: 2,259 Days
• Field of Search: 382/185, 382/187, 382/229, 382/161, 382/186, 382/188, 382/189
• US Class Current: 382/187
• CPC Class Codes: G06V 10/75 Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries

Abstract

A pattern recognizing apparatus which recognizes patterns on the basis of a recognition dictionary includes a frequency information updating device for updating a plurality of items of frequency information, respectively assigned to a plurality of reference patterns, in accordance with a reference to at least one of the plurality of reference patterns during recognition and a recognition dictionary reconfiguration device for reconfiguring the recognition dictionary based on a difference between the updated frequency information.

Claims

1. A pattern recognizing apparatus comprising:

candidate display control means for controlling a display of a plurality of candidate characters obtained by recognizing one input pattern;

selection means for selecting one candidate character among the displayed candidate characters in accordance with a user'"'"'s selection;

frequency information updating means for updating a frequency information of reference patterns which is stored in a recognition dictionary, in relation to the user'"'"'s selection of said one candidate character; and

recognition dictionary reconfiguration means for reconfiguring an order of reference patterns stored in said recognition dictionary on the basis of a relationship between the updated frequency information.

2. A pattern recognizing apparatus according to claim 1, further comprising user registration means which, in a case that a reference pattern is added through user registration, adds the reference pattern to the recognition dictionary as a reference pattern of high order of reference.

3. A pattern recognizing apparatus according to claim 1, wherein each time said reference patterns are referred to, said frequency information updating means increases the frequency information of a corresponding item of frequency information.

4. A pattern recognizing apparatus according to claim 3, wherein, when at least one of the differences of said frequency information has reached a predetermined value, said recognition dictionary reconfiguration means reconfigures said recognition dictionary.

5. A pattern recognizing apparatus according to one of claim 1, wherein said recognition dictionary reconfiguration means reconfigures said recognition dictionary at a predetermined point in time.

6. A pattern recognizing apparatus according to claim 1, wherein said recognition dictionary reconfiguration means performs rearrangement of reference patterns according to an order of frequency of reference, determined from said frequency information, so as to initialize the frequency information.

7. A pattern recognizing apparatus according to claim 6, wherein, when said frequency of reference is not higher than a predetermined frequency, said recognition dictionary reconfiguration means deletes a corresponding reference pattern.

8. A pattern recognizing apparatus according to claim 1, wherein said recognition dictionary reconfiguration means checks whether a pattern is one which is repeatedly referred to even after a predetermined period of time has elapsed, and which, in a case that a pattern has been generated that is repeatedly referred to at the predetermined frequency after the lapse of the predetermined period of time, reconfigures said recognition dictionary.

9. A pattern recognizing apparatus according to claim 4, wherein said recognition dictionary reconfiguration means performs reconfiguration according to an order of frequency of reference, determined from said frequency information, so as to initialize the frequency information.

10. A pattern recognizing apparatus according to claim 9, wherein, when said frequency of reference is not higher than a predetermined frequency, said recognition dictionary reconfiguration means deletes a corresponding reference pattern.

11. A pattern apparatus according to claim 2, wherein said user registration means comprises registered pattern display means for displaying registered patterns and registered pattern information rearranging means for rearranging registered patterns displayed according to a predetermined condition.

12. A pattern recognizing apparatus according to claim 11, wherein said registered pattern information rearranging means rearranges registered patterns according to a number of strokes of each of the registered patterns.

13. A pattern recognizing apparatus according to claim 11, wherein said registered pattern information rearranging means rearranges registered patterns according to dates of registration of the registered patterns.

14. A pattern recognizing apparatus according to claim 11, wherein said registered pattern information rearranging means rearranges registered patterns according to radical information of the registered patterns.

29. A pattern recognizing apparatus according to claim 1, wherein said relationship between the updated frequency information is calculated from updated frequency information of a plurality of reference patterns which are included in the same category and which have a different sequence of strokes from each other.

15. Computer-executable process steps stored on a computer-readable medium, the steps to recognize patterns on the basis of a recognition dictionary, the steps comprising:

a receiving step to receive a user'"'"'s selection of one candidate character among a plurality of displayed candidate characters obtained by recognizing one input pattern;

an update step to update a frequency information assigned respectively to a plurality of reference patterns on the basis of the user'"'"'s selection of said one candidate character;

a reconfiguring step to reconfigure an order of reference patterns stored in the recognition dictionary on the basis of a relationship between the updated frequency information assigned to respective reference patterns; and

a recognizing step to recognize patterns on the basis of the reconfigured recognition dictionary.

16. Computer-executable process steps according to claim 15, wherein, in said updating step, each time the reference patterns are referred to, corresponding frequency information is increased, and

wherein, in said reconfiguring step, the reconfiguration of said recognition dictionary is executed when at least one item of said frequency information has reached a predetermined value.

17. Computer-executable process steps according to claim 16, wherein, in said reconfiguring step, the reconfiguration is effected in an order of frequency of reference, determined from said frequency information, so as to reconfigure frequency information.

18. Computer-executable process steps according to claim 16, wherein, in said reconfiguring step, in a case that said frequency of reference is not higher than a predetermined level, a corresponding reference pattern is deleted.

19. Computer-executable process steps according to claim 17, wherein, in said reconfiguring step, in a case that said frequency of reference is not higher than a predetermined level, a corresponding reference pattern is deleted.

20. Computer-executable process steps according to claim 15, wherein, in said reconfiguring step, said recognition dictionary reconfiguration means checks whether a pattern is one which is repeatedly referred to even after a predetermined period of time has elapsed, and which, in a case that a pattern has been generated which is repeatedly referred to at a predetermined frequency after the lapse of the predetermined period of time, reconfigures said recognition dictionary.

21. Computer-executable process steps according to claim 15, wherein said relationship between the updated frequency information is calculated from updated frequency information of a plurality of reference patterns which are included in the same category and which have a different sequence of strokes from each other.

22. A pattern recognizing method in which patterns are recognized on the basis of a recognition dictionary, said pattern recognizing method comprising the steps of:

displaying a plurality of candidate characters obtained by recognizing one input pattern;

updating a frequency information assigned respectively to a plurality of reference patterns on the basis of a user'"'"'s selection of one candidate character among the displayed candidate characters;

reconfiguring an order of reference patterns stored in the recognition dictionary on the basis of a relationship between the updated frequency information; and

recognizing patterns on the basis of the reconfigured recognition dictionary.

23. A pattern recognizing method according to claim 22, wherein said relationship between the updated frequency information is calculated from updated frequency information of a plurality of reference patterns which are included in the same category and which have a different sequence of strokes from each other.

24. A pattern recognizing method according to claim 22,

wherein, in said step of updating the frequency information, each time the reference patterns are referred to, corresponding frequency information is increased, and

wherein, in said step of reconfiguring said recognition dictionary, the reconfiguration of said recognition dictionary is executed when at least one item of said frequency information has reached a predetermined value.

25. A pattern recognizing method according to claim 24, wherein, in said step of reconfiguring said recognition dictionary, the reconfiguration is effected in an order of frequency of reference, determined from said frequency information, so as to reconfigure frequency information.

26. A pattern recognizing method according to claim 24, wherein, in said step of reconfiguring said recognition dictionary, in a case that said frequency of reference is not higher than a predetermined level, a corresponding reference pattern is deleted.

27. A pattern recognizing method according to claim 25, wherein, in said step of reconfiguring said recognition dictionary, in a case that said frequency of reference is not higher than a predetermined level, a corresponding reference pattern is deleted.

28. A pattern recognizing method according to claim 22, wherein, in said step of reconfiguring the recognition dictionary, said recognition dictionary reconfiguration means checks whether a pattern is one which is repeatedly referred to even after a predetermined period of time has elapsed, and which, in a case that a pattern has been generated which is repeatedly referred to at a predetermined frequency after the lapse of the predetermined period of time, reconfigures said recognition dictionary.