Device, method and storage medium for recognizing a document image

• US 6,347,156 B1
• Filed: 12/21/1998
• Issued: 02/12/2002
• Est. Priority Date: 05/27/1998
• Status: Expired due to Term

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Kamada, Hiroshi, Fujimoto, Katsuhito
• Primary Examiner(s): Boudreau, Leo
• Assistant Examiner(s): Mariam, Daniel G.
• Application Number:
  US09/216,712
• Time in Patent Office: 1,149 Days
• Field of Search: 382/162, 382/164, 382/165, 382/166, 382/169, 382/173, 382/176, 382/181, 382/190, 382/195, 382/237, 382/299, 382/321, 382/205, 382/232, 382/300, 358/448, 358/455, 358/500, 358/429
• US Class Current: 382/237
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/162 Quantising the image signal
• CPC Class Codes: H04N 1/40012 Conversion of colour to monochrome
• CPC Class Codes: H04N 1/40068 Modification of image resolution, i.e. determining the values of picture elements at new relative positions H04N1/3935 takes precedence

Abstract

A color image input from a document image inputting unit is converted into a gray-scale image by a brightness image extracting unit. The gray-scale image is then converted into an image having a higher resolution according to the resolution of the original gray-scale image. When this conversion is performed, subpixels are generated between the original pixels, and the values of the subpixels are obtained with an interpolation method. Furthermore, a threshold value for a binarization process is generated by using an original pixel value and a subpixel value. The characters included in the binarized image are recognized by a binary image recognizing unit, and a recognition result is output from a recognition result outputting unit.

Claims

1. A document image recognizing device, comprising:

image converting means for converting an input document image into a gray-scale image if the input document image is a color image, and for newly outputting a gray-scale image if the input image is a gray-scale image;

variable resolution binarizing means for converting the input document image into a binary image having a higher resolution according to a resolution of the gray-scale image;

wherein said variable resolution binarizing means performs a sub-pixel generation process for increasing a number of pixels included in an image by interpolating pixel values ofa gray-scale image, sets a local threshold value within a local area centering around a particular pixel, and obtains a binary image by using the local threshold value; and

recognizing means for recognizing the binarized image.

2. The device according to claim 1, wherein:

said recognizing means recognizes a converted or input binary image, and converts the binary image into electronic codes.

3. The device according to claim 1, further comprising:

drawing area roughly extracting means for roughly extracting a drawing area according to a global threshold value for pixel values of a gray-scale image, wherein

said recognizing means recognizes a binary image in an area extracted by said drawing area roughly extracting means.

4. The device according to claim 3, wherein:

the global threshold value uses a linear combination of an average pixel value, a standard deviation value, and a variance.

5. The device according to claim 3, wherein

said variable resolution binarizing means performs the subpixel generation process for an entire input gray-scale image;

and

said drawing area roughly extracting means roughly extracts a drawing area from gray-scale image data for which the subpixel generation process is performed.

6. The device according to claim 1, wherein:

the subpixel generation process is performed by linearly interpolating original pixel values of the gray-scale image.

7. The device according to claim 1, wherein:

said variable resolution binarizing means obtains a binary image by binarizing pixel values by using the local threshold value obtained from a distribution of the pixel values within a local area including pixels; and

the local threshold value uses a linear combination of an average pixel value, a standard deviation value, and a variance.

8. The device according to claim 1, further comprising:

drawing area roughly extracting means for performing a global process which roughly extracts a drawing area according to a global threshold value for pixel values of a gray-scale image, wherein

said variable resolution binarizing means performs the subpixel generation process for increasing a number of pixels included in an image by interpolating pixel values of a gray-scale image, for a drawing area which is roughly extracted with the global process, and performs a local binarization process by using the local threshold value for each pixel included in the roughly extracted drawing area.

9. The device according to claim 1, wherein the local threshold value at a pixel point after the subpixel generation process is based on obtaining a value of a subpixel generated with the subpixel generation process from original pixel values, and obtaining a local threshold value from subpixel values.

10. The device according to claim 1, wherein a local threshold value after the subpixel generation process is obtained at a pixel point of a gray-scale image, and a local threshold value of a subpixel is obtained by interpolating local threshold values at pixel points of the gray-scale image.

11. The device according to claim 10, wherein the local threshold value at the pixel point of the gray-scale image is obtained by using a value of a subpixel obtained by performing the subpixel generation process for pixel points of the gray-scale image.

12. The device according to claim 10, wherein the local threshold value at the pixel point of the gray-scale image after the subpixel generation process is based on obtaining a subpixel value from original pixel values, and obtaining a local threshold value from subpixel values.

13. The device according to claim 12, wherein the local threshold value is obtained, using subpixel values, from a local threshold value from a local area that recognizes original pixels as its boundary.

14. The device according to claim 10, wherein an interpolation value at a subpixel point is obtained by interpolating a difference between a value at a pixel point of a gray-scale image and a local threshold value after the subpixel generation process, and a value of a binary image at the subpixel point is determined with a sign of the interpolation value.

15. The device according to claim 1, wherein a local threshold value at an original pixel point of a gray-scale image after the subpixel generation process is obtained from a linear combination of an average value, a standard deviation value, and a variance, by obtaining the standard deviation value and the variance after obtaining the average value and a square average value of values at pixel points of the gray-scale image, for which the subpixel generation process is performed, based on obtaining a subpixel value from original pixel values, and obtaining the average value and the square average value from subpixel values.

16. The device according to claim 1, wherein:

the local area recognizes original pixels as its boundary; and

a local threshold value at an original pixel point of a gray-scale image after the subpixel generation process is obtained from a linear combination of an average value, a standard deviation value, and a variance, by obtaining the standard deviation value and the variance after obtaining the average value and a square average value of values at pixel points of the gray-scale image, for which the subpixel generation process is performed, based on obtaining a subpixel value from original pixel values, and obtaining the average value and the square average value from subpixel values.

17. The device according to claim 1, further comprising:

specifying means for specifying a range of a local area, wherein

whether or not the local area recognizes original pixels as its boundary is determined by using as specification data a number of subpixels generated with the subpixel generation process and a size of the local area.

18. The device according to claim 1, wherein:

a local threshold value after the subpixel generation process is obtained at a pixel point of a gray-scale image;

a table to which pixel values and local threshold values at four pixel points of a gray-scale are input and from which a binary image enclosed by the four pixel points is output as data, is included; and

a binary image is obtained by using said table.

19. A document image recognizing method, comprising:

converting an input document image into a gray-scale image if the input document image is a color image, and newly outputting a gray-scale image if the input image is a gray-scale image;

converting the input document image into a binary image having a higher resolution according to a resolution of the gray-scale image;

wherein said converting the input document image into a binary image performs a sub-pixel generation process for increasing a number of pixels included in an image by interpolating pixel values of a gray-scale image, sets a local threshold value within a local area centering around a particular pixel, and obtains a binary image by using the local threshold value; and

recognizing the binary image.

20. The method according to claim 19, wherein:

said recognizing the binary image recognizes a converted or input binary image, and converts the binary image into electronic codes.

21. The method according to claim 20, further comprising:

roughly extracting a drawing area according to a global threshold value for pixel values of a gray-scale image, wherein

said recognizing the binary image recognizes a binary image in an area extracted by said roughly extracting a drawing area.

22. The method according to claim 21, wherein the global threshold value uses a linear combination of an average pixel value, a standard deviation value, and a variance.

23. The method according to claim 21, wherein:

said converting the input document image into a binary image performs the subpixel generation process for an entire input gray-scale image; and

said roughly extracting a drawing area roughly extracts a drawing area from gray-scale image data for which the subpixel generation process is performed.

24. The method according to claim 19, wherein:

the subpixel generation process is performed by linearly interpolating original pixel values of the gray-scale image.

25. The method according to claim 19, wherein:

said converting the input document image into a binary image obtains a binary image by binarizing pixel values by using the local threshold value obtained from a distribution of the pixel values within a local area including pixels; and

the local threshold value uses a linear combination of an average pixel value, a standard deviation value, and a variance.

26. The method according to claim 19, further comprising:

performing a global process which roughly extracts a drawing area according to a global threshold value for pixel values of a gray-scale image, wherein

said converting the input document image into a binary image performs the subpixel generation process for increasing a number of pixels included in an image by interpolating pixel values of a gray-scale image, for a drawing area which is roughly extracted with the global process, and performs a local binarization process by using the local threshold value for each pixel included in the roughly extracted drawing area.

27. The method according to claim 19, wherein the local threshold value at a pixel point after the subpixel generation process is based on obtaining a value of a subpixel generated with the subpixel generation process from original pixel values, and obtaining a local threshold value from subpixel values.

28. The method according to claim 19, wherein a local threshold value after the subpixel generation process is obtained at a pixel point of a gray-scale image, and a local threshold value of a subpixel is obtained by interpolating local threshold values at pixel points of the gray-scale image.

29. The method according to claim 28, wherein the local threshold value at the pixel point of the gray-scale image after the subpixel generation process is based on obtaining a subpixel value from original pixel values, and obtaining a local threshold value from subpixel values.

30. The method according to claim 29, wherein the local threshold value is obtained, using subpixel values, from a local threshold value within a local area that recognizes original pixels as its boundary.

31. The method according to claim 28, wherein an interpolation value at a subpixel point is obtained by interpolating a difference between a value at a pixel point of a gray-scale image and a local threshold value after the subpixel generation process, and a value of a binary image at the subpixel point is determined by a sign of the interpolation value.

32. The method according to claim 19, wherein the local threshold value at the pixel point of the gray-scale image is obtained by using a value of a subpixel obtained by performing the subpixel generation process for pixel points of the gray-scale image.

33. The method according to claim 19, wherein a local threshold value at an original pixel point of a gray-scale image after the subpixel generation process is obtained from a linear combination of an average value, a standard deviation value, and a variance, by obtaining the standard deviation value and the variance after obtaining the average value and a square average value of values at pixel points of the gray-scale image, for which the subpixel generation process is performed, based on obtaining a subpixel value from original pixel values, and obtaining the average value and the square average value from subpixel values.

34. The method according to claim 19, wherein:

the local area recognizes original pixels as its boundary; and

a local threshold value at an original pixel point of a gray-scale image after the subpixel generation process is obtained from a linear combination of an average value, a standard deviation value, and a variance, by obtaining the standard deviation value and the variance after obtaining the average value and a square average value of values at pixel points of the gray-scale image, for which the subpixel generation process is performed, based on obtaining a subpixel value from original pixel values, and obtaining the average value and the square average value from subpixel values.

35. The method according to claim 19, further comprising:

specifying a range of a local area, wherein whether or not the local area recognizes original pixels as its boundary is determined by using as specification data a number of subpixels generated with the subpixel generation process and a size of the local area.

36. The method according to claim 19, wherein:

a local threshold value after the subpixel generation process is obtained at a pixel point of a gray-scale image;

a table to which pixel values and local threshold values at four pixel points of a gray-scale are input and from which a binary image enclosed by the four pixel points is output as data, is included; and

a binary image is obtained by using said table.

37. A computer-readable storage medium for directing a computer to execute a process comprising:

converting an input document image into a gray-scale image if the input document image is a color image, and newly outputting a gray-scale image if the input image is a gray-scale image;

converting the input document image into a binary image having a higher resolution according to a resolution of a gray-scale image;

wherein said converting the input document image into a binary image performs a sub-pixel generation process for increasing a number of pixels included in an image by interpolating pixel values of a gray-scale image, sets a local threshold value within a local area centering around a particular pixel, and obtains a binary image by using the local threshold value; and

recognizing the binary image.

38. The storage medium according to claim 37, wherein:

said recognizing the binary image recognizes a converted or input binary image, and converts the binary image into electronic codes.

39. The storage medium according to claim 37, wherein the process further comprises:

roughly extracting a drawing area according to a global threshold value for pixel values of a gray-scale image, and wherein

said recognizing the binary image recognizes a binary image in an area extracted by said roughly extracting a drawing area.

40. The storage medium according to claim 39, wherein the global threshold value uses a linear combination of an average pixel value, a standard deviation value, and a variance.

41. The storage medium according to claim 39, wherein:

said converting the input document image into a binary image performs the subpixel generation process for an entire input gray-scale image; and

said roughly extracting a drawing area roughly extracts a drawing area from gray-scale image data for which the subpixel generation process is performed.

42. The storage medium according to claim 39, wherein the subpixel generation process is performed by linearly interpolating original pixel values of the gray-scale image.

43. The storage medium according to claim 37, wherein:

said converting the input document image into a binary image obtains a binary image by binarizing pixel values by using the local threshold value obtained from a distribution of the pixel values within a local area including pixels; and

the local threshold value uses a linear combination of an average pixel value, a standard deviation value, and a variance.

44. The storage medium according to claim 37, further comprising:

performing a global process which roughly extracts a drawing area according to a global threshold value for pixel values of a gray-scale image, and wherein

said converting the input document image into a binary image performs the subpixel generation process for increasing a number of pixels included in an image by interpolating pixel values of a gray-scale image, for a drawing area which is roughly extracted with the global process, and performs a local binarization process by using the local threshold value for each pixel included in the roughly extracted drawing area.

45. The storage medium according to claim 37, wherein the local threshold value at a pixel point after the subpixel generation process is based on obtaining a value of a subpixel generated with the subpixel generation process from original pixel values, and obtaining a local threshold value from subpixel values.

46. The storage medium according to claim 37, wherein a local threshold value after the subpixel generation process is obtained at a pixel point of a gray-scale image, and a local threshold value of a subpixel is obtained by interpolating local threshold values at pixel points of the gray-scale image.

47. The storage medium according to claim 46, wherein the local threshold value at the pixel point of the gray-scale image is obtained by using a value of a subpixel obtained by performing the subpixel generation process for pixel points of the gray-scale image.

48. The storage medium according to claim 46, wherein the local threshold value at the pixel point of the gray-scale image after the subpixel generation process is based on obtaining a subpixel value from original pixel values, and obtaining a local threshold value from subpixel values.

49. The storage medium according to claim 48, wherein the local threshold value is obtained, using subpixel values, from a local threshold value within a local area that recognizes original pixels as its boundary.

50. The storage medium according to claim 46, wherein an interpolation value at a subpixel point is obtained by interpolating a difference between a value at a pixel point of a gray-scale image and a local threshold value after the subpixel generation process, and a value of a binary image at the subpixel point is determined with a sign of the interpolation value.

51. The storage medium according to claim 37, wherein a local threshold value at an original pixel point of a gray-scale image after the subpixel generation process is obtained from a linear combination of an average value, a standard deviation value, and a variance, by obtaining the standard deviation value and the variance after obtaining the average value and a square average value of values at pixel points of the gray-scale image, for which the subpixel generation process is performed, based on obtaining a subpixel value from original pixel values, and obtaining the average value and the square average value from subpixel values.

52. The storage medium according to claim 37, wherein:

the local area recognizes original pixels as its boundary; and

a local threshold value at an original pixel point of a gray-scale image after the subpixel generation process is obtained from a linear combination of an average value, a standard deviation value, and a variance, by obtaining the standard deviation value and the variance after obtaining the average value and a square average value of values at pixel points of the gray-scale image, for which the subpixel generation process is performed, based on obtaining a subpixel value from original pixel values, and obtaining the average value and the square average value from subpixel values.

53. The storage medium according to claim 37, further comprising:

specifying a range of a local area, and wherein

whether or not the local area recognizes original pixels as its boundary is determined by using as specification data a number of subpixels generated with the subpixel generation process and a size of the local area.

54. The storage medium according to claim 37, wherein:

a local threshold value after the subpixel generation process is obtained at a pixel point of a gray-scale image;

a table to which pixel values and local threshold values at four pixel points of a gray-scale are input and from which a binary image enclosed by the four pixel points is output as data, is included; and

a binary image is obtained by using said table.