Character input device, character input method and character input program

• US 20030169924A1
• Filed: 03/07/2003
• Published: 09/11/2003
• Est. Priority Date: 03/08/2002
• Status: Active Grant

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Senda, Shuji, Nishiyama, Kyosuke
• Granted Patent: US 7,221,796 B2
• Time in Patent Office: Days
• US Class Current: 382/182
• CPC Class Codes: G06V 30/142 using hand-held instruments; Constructional details of the instruments

Abstract

A character input device including a data processing device which is provided with a character extraction device for extracting an image of a character portion from an image photographed by a photographing device, a character recognition device for recognizing a character from the image of the character portion and making a character recognition result with its reliability added into character recognition data, and a recognition data synthesizing device for detecting and connecting overlap portions in the character recognition data obtained from a plurality of images of the same document, as well as selecting character recognition data whose reliability is high with respect to the overlap portions to synthesize the plurality of the character recognition data into one synthetic recognition data; a recognition data storage device for storing synthetic recognition data; and a display device for displaying a photographed image and a character recognition result.

Claims

1. A character input device for inputting a character by extracting and recognizing a character from an image of a document photographed by a photographing device, wherein

a recognition result of said character is stored with reliability indicative of sureness of recognition added, and

character recognition results of images obtained by repeatedly photographing the whole or a part of said document are compared to input a character based on a character recognition result whose said reliability is the highest.

2. The character input device as set forth in claim 1, wherein

said reliability is determined based on at least information about a degree of coincidence between a dictionary to be referred in character recognition and a character extracted from said photographed image and about distortion of said photographed image.

9. The character input device as set forth in claim 1, further comprising

means for detecting an amount of movement of said photographing device from a difference between successively input images and discriminating a stationary state of said photographing device based on said amount of movement to automatically give an image photographing instruction.

3. A character input device for inputting a character by extracting and recognizing a character from an image of a document photographed by a photographing device, comprising:

character extraction means for extracting an image of a character portion from an image photographed by said photographing device;

character recognition means for recognizing a character from the image of said character portion and making a character recognition result with said reliability added into character recognition data;

recognition data synthesizing means for detecting and connecting overlap portions in said character recognition data obtained from a plurality of images of the same document, as well as selecting character recognition data whose said reliability is high with respect to the overlap portions to synthesize the plurality of said character recognition data into one said character recognition data; and

recognition data storage means for storing said character recognition data synthesized.

4. The character input device as set forth in claim 3, wherein

said reliability is determined based on at least information about a degree of coincidence between a dictionary to be referred in character recognition and a character extracted from said photographed image and about distortion of said photographed image.

5. The character input device as set forth in claim 3, wherein

said recognition data synthesizing means detects,

with respect to said character recognition data obtained from a plurality of images of the same document, an overlap portion between said character recognition data obtained from the latest image and said character recognition data obtained from a former image which is stored in said recognition data storage means and when said reliability of said character recognition data obtained from said latest image is higher, rewrites said character recognition data obtained from the former image in said recognition data storage means.

6. The character input device as set forth in claim 3, further comprising

character correction means for reading, from said recognition data storage means, at least one other recognition candidate for a character designated from among characters stored as character recognition data and presenting the read candidate, and replacing character recognition data of said designated character with character recognition data of a character selected from among said recognition candidates to correct said character recognition data.

7. The character input device as set forth in claim 6, wherein

said character correction means selects,

from among said character recognition data displayed on display means, character recognition data of a character located in the vicinity of the center of a photographed image as a character to be corrected.

8. The character input device as set forth in claim 3, wherein

said recognition data synthesizing means detects,

with respect to said character recognition data obtained from a plurality of images of the same document, an overlap portion between said character recognition data obtained from the latest image and said character recognition data obtained from a former image which is stored in said recognition data storage means and when none of said overlap portion is detected, stores said character recognition data obtained from said latest image as pending data and detects an overlap portion of character recognition data obtained from a subsequently input image with said pending data.

10. The character input device as set forth in claim 3, further comprising:

image temporary storage means for temporarily storing an image photographed by said photographing device, and

image difference detecting means for detecting a difference between an image stored in said image temporary storage means and an image input from said photographing device and when said difference detected is not more than a predetermined threshold value, determining that said photographing device is at a stationary state to send a photographing instruction to said photographing device.

11. The character input device as set forth in claim 10, wherein

said image difference detecting means determines that said photographing device is at the stationary state when a state where said difference is not more than the predetermined threshold value continues for a predetermined period of time.

31. The character input device as set forth in claim 3, further comprising:

mark drawing means for drawing a mark indicative of a position of a specific character string designated in an image photographed by said photographing device, and

line extraction means for dividing said photographed image into line images on a line basis and sequentially sending the divisional line images to said character extraction means starting with a line image closest to said mark.

32. The character input device as set forth in claim 3, wherein

said recognition data synthesizing means includes

forced connection means for, when no overlap portion is detected between said character recognition data obtained from the latest image and said character recognition data obtained from a former image which is stored in said recognition data storage means, forcedly connecting said character recognition data according to an instruction from said user.

33. The character input device as set forth in claim 31, further comprising:

position control means for determining whether data is character recognition data from an initially photographed image based on character recognition data stored in said recognition data storage means and indicating the position of the top of a character string to be recognized when the character recognition data is of the initial photographed image and otherwise indicating the position of the end of the character string to be recognized, and

deletion mans for deleting a character of said character recognition data based on an indicated position at the top or the end of a character string to be recognized.

12. A character input method of inputting a character by extracting and recognizing a character from an image of a document photographed by a photographing device, comprising the steps of:

adding reliability indicative of sureness of recognition to a recognition result of said character;

comparing character recognition results of images obtained by repeatedly photographing the whole or a part of said document; and

selecting a character recognition result whose said reliability is the highest.

13. The character input method as set forth in claim 12, wherein

said reliability is determined based on at least information about a degree of coincidence between a dictionary to be referred in character recognition and a character extracted from said photographed image and about distortion of said photographed image.

14. A character input method as set forth in claim 12, further comprising the steps of:

detecting and connecting overlap portions in said character recognition data obtained from a plurality of images of the same document, and

selecting character recognition data whose said reliability is high with respect to the overlap portions to synthesize the plurality of said character recognition data into one said character recognition data.

15. The character input method as set forth in claim 14, further comprising the step of,

in said recognition data synthesis, with respect to said character recognition data obtained from a plurality of images of the same document, detecting an overlap portion between said character recognition data obtained from the latest image and said character recognition data obtained from a former image and when said reliability of said character recognition data obtained from said latest image is higher, rewriting said character recognition data obtained from the former image.

16. The character input method as set forth in claim 14, further comprising the step of

designating a character stored as character recognition data to present other candidates for the designated character and replacing character recognition data of said designated character with character recognition data of a character selected from among said candidates to correct said character recognition data.

17. The character input method as set forth in claim 16, further comprising the step of

moving a cursor to character recognition data of a character located in the vicinity of the center of a photographed image among said character recognition data displayed to indicate that it is highly probable that the character will be corrected.

18. The character input method as set forth in claim 14, further comprising the step of

detecting, in said recognition data synthesis, with respect to said character recognition data obtained from a plurality of images of the same document, an overlap portion between said character recognition data obtained from the latest image and said character recognition data obtained from a former image and when detecting none of said overlap portion, storing said character recognition data obtained from said latest image as pending data and detecting an overlap portion of character recognition data obtained from a subsequently input image with said pending data.

19. The character input method as set forth in claim 12, further comprising step of

detecting an amount of movement of said photographing device from a difference between successively input images and discriminating a stationary state of said photographing device based on said amount of movement to automatically execute image photographing.

20. The character input method as set forth in claim 14, further comprising the steps of:

temporarily storing an image photographed by said photographing device, and

detecting a difference between said temporarily stored image and an image input from said photographing device and when said difference detected is not more than a predetermined threshold value, determining that said photographing device is at a stationary state to send a photographing instruction to said photographing device.

21. The character input method as set forth in claim 20, further comprising the step of

determining that said photographing device is at the stationary state when a state where said difference is not more than the predetermined threshold value continues for a predetermined period of time.

22. A character input program executed on a computer for inputting a character by extracting and recognizing a character from an image of a document photographed by a photographing device, comprising the functions of:

adding reliability indicative of sureness of recognition to a recognition result of said character, and

comparing character recognition results of images obtained by repeatedly photographing the whole or a part of said document to select a character recognition result whose said reliability is the highest.

23. The character input program as set forth in claim 22, wherein

said reliability is determined based on at least information about a degree of coincidence between a dictionary to be referred in character recognition and a character extracted from said photographed image and about distortion of said photographed image.

24. The character input program as set forth in claim 22, further comprising the functions of:

detecting and connecting overlap portions in said character recognition data obtained from a plurality of images of the same document, and

selecting character recognition data whose said reliability is high with respect to the overlap portions to synthesize the plurality of said character recognition data into one said character recognition data.

25. The character input program as set forth in claim 24, wherein

said recognition data synthesis functions includes

detecting, with respect to said character recognition data obtained from a plurality of images of the same document, an overlap portion between said character recognition data obtained from the latest image and said character recognition data obtained from a former image and when said reliability of said character recognition data obtained from said latest image is higher, rewriting said character recognition data obtained from the former image.

26. The character input program as set forth in claim 24, further comprising the function of

the character correction function of designating a character stored as character recognition data to present other candidates for the designated character and replacing character recognition data of said designated character with character recognition data of a character selected from among said candidates to correct said character recognition data.

27. The character input program as set forth in claim 26, wherein

the character correction function includes moving a cursor to character recognition data of a character located in the vicinity of the center of a photographed image among said character recognition data displayed to indicate that it is highly probable that the character will be corrected.

28. The character input program as set forth in claim 24, wherein

said recognition data synthesis function includes

detecting, with respect to said character recognition data obtained from a plurality of images of the same document, an overlap portion between said character recognition data obtained from the latest image and said character recognition data obtained from a former image and when detecting none of said overlap portion, storing said character recognition data obtained from said latest image as pending data and detecting an overlap portion of character recognition data obtained from a subsequently input image with said pending data.

29. The character input program as set forth in claim 22, further comprising the function of

a function of detecting an amount of movement of said photographing device from a difference between successively input images and discriminating a stationary state of said photographing device based on said amount of movement to automatically execute image photographing.

30. The character input program as set forth in claim 24, further comprising the functions of:

temporarily storing an image photographed by said photographing device, and

detecting a difference between said temporarily stored image and an image input from said photographing device and when said difference detected is not more than a predetermined threshold value, determining that said photographing device is at a stationary state to send a photographing instruction to said photographing device.