Image processing method and apparatus employing gray scale images having fewer bits per pixel and gray scale image restoration using small areas of an image window

• US 5,521,987 A
• Filed: 06/01/1994
• Issued: 05/28/1996
• Est. Priority Date: 06/04/1993
• Status: Expired due to Term

• Current Assignee: Omron Corporation
• Original Assignee: Omron Corporation
• Inventors: Masaki, Toshimichi
• Primary Examiner(s): Boudreau, Leo
• Assistant Examiner(s): Johns, Andrew W.
• Application Number:
  US08/251,966
• Time in Patent Office: 727 Days
• Field of Search: 382/209, 382/216, 382/218, 382/237, 382/251, 382/252, 382/270, 382/296, 382/298, 382/299, 358/456, 358/457, 348/161
• US Class Current: 382/218
• CPC Class Codes: G06T 3/40 Scaling of whole images or parts thereof, e.g. expanding or contracting
• CPC Class Codes: G06V 10/751 Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
• CPC Class Codes: G06V 2201/02 Recognising information on displays, dials, clocks

Abstract

Input image data representing the image of a subject imaged by a video camera 9 is applied from an A/D converter circuit 11 to a dither-image generating circuit 13, in which the image data is binarized in accordance with the error diffusion method, whereby dither image data represented by one bit per pixel is obtained. A plurality of small areas each comprising a plurality of pixels are set within a window caused to scan the dither image, and the dither image data is summed in each and every small area by an image restoration circuit 14, whereby gray-level image data, in which each small area serves as one pixel, is restored. Difference or correlation values between the restored image data and text data set in advance with regard to standard images which are different from one another are calculated, for the window at each and every position, by arithmetic circuits 15a˜15k. The smallest difference or the largest correlation values and the positions of the windows in which this smallest difference or the largest correlation values are produced are respectively sensed by corresponding peak holding circuits 16a˜16k.

Claims

1. An image processor comprising:

pseudo-gray-scale image generating circuit means for converting input image data, which is represented by a first prescribed number of bits per pixel, into pseudo-gray-scale image data represented by a second prescribed number of bits, which is less than said first prescribed number of bits;

image restoration circuit means for restoring the pseudo-gray-scale image data generated by said pseudo-gray-scale image generating circuit means for each and every one of a plurality of small areas, each of which comprises a plurality of pixels, with said each small area serving as one restored pixel, and outputting restored image data; and

a plurality of similarity operating circuit means, which operate in parallel, each for calculating an index value representing a degree of similarity between the restored image data outputted by said image restoration circuit means and text data representing a standard image set in advance, the standard images for the plurality of similarity operating circuit means being different from one another.

2. An image processor according to claim 1, wherein the standard images are images of models which are different from one another.

3. An image processor according to claim 1, wherein the standard images are formed based on a single model and different in rotational angle or in size from one another.

4. An image processor according to claim 1 further comprising judging means for determining whether a subject image represented by the input image data coincides with at least one of the standard images on the basis of the index values calculated by the plurality of said similarity operating circuit means.

5. An image processor according to claim 1, wherein said pseudo-gray-scale image generating circuit means binarizes the input image data using an error diffusion method and generates dither image data represented by one bit per pixel.

6. An image processor according to claim 1, wherein said pseudo-gray-scale image generating circuit means subjects the input image data to a four-value conversion using an error diffusion method and generates pseudo-gray-scale image data represented by two bits per pixel.

7. An image processor according to claim 1, wherein

said image restoration circuit means restores the pseudo-gray-scale image data for each and every one of a plurality of small areas, each of which is provided in a window that scans the pseudo-gray-scale image generated by said pseudo-gray-scale image generating circuit means, and outputting restored image data for a window at each and every position;

and

said similarity operating circuit means calculates the index value representing a degree of similarity between the restored image data of each window sequentially outputted by the image restoration circuit means and text data representing a standard image set in advance.

8. An image processor according to claim 7, further comprising a plurality of peak holding circuit means each for detecting an index value, which represents the highest degree of similarity, from among similarity index values regarding each of the windows obtained from said corresponding similarity operating circuit means, and position data indicative of the window in which this index value is generated.

9. An image processor according to claim 8, further comprising judging means for determining whether a subject image represented by the input image data has an image portion which coincides with at least one of the standard images on the basis of the index values detected by the plurality of said peak holding circuit means.

10. An image processor according to claim 7, wherein the index value representing said degree of similarity is a difference value in lightness of the image, and said similarity operating circuit means calculates an absolute value of a difference between restored image data and corresponding text data for each and every small area, sums the absolute values of these differences over all small areas contained in one window and outputs the sum value as a difference value of lightness.

11. An image processor according to claim 10, further comprising a peak holding circuit means for detecting a smallest difference value from among difference values of lightness regarding each of the windows obtained by said similarity operating circuit means, as well as position data indicative of the window in which the smallest difference value is produced.

12. An image processor according to claim 7, wherein said similarity operating circuit means calculates a cross-correlation value between restored image data and text data for each and every window, and outputs this cross-correlation value as the index value representing degree of similarity.

13. An image processor according to claim 12, further comprising a peak holding circuit means for detecting a largest cross-correlation value from among the cross-correlation values regarding each of the windows obtained by said similarity operating circuit means, as well as position data indicative of the window in which the largest cross-correlation value is produced.

14. An image processor comprising:

converting means for converting input image data into image data representing images different in rotational angle or in size from the image represented by the input image data in sequence at a predetermined period of time;

pseudo-gray-scale image generating circuit means for converting the image data converted by said converting means and represented by a first prescribed number of bits per pixel, every time the converted image data is supplied from said converting means, into pseudo-gray-scale image data represented by a second prescribed number of bits, which is less than said first prescribed number of bits;

image restoration circuit means for restoring the pseudo-gray-scale image data for each and every one of a plurality of small areas, each of which comprises a plurality of pixels, with said each small area serving as one restored pixel, every time the pseudo-gray-scale image data is supplied from said pseudo-gray-scale image generating circuit means and outputting restored image data; and

similarity operating circuit means for calculating an index value representing a degree of similarity between the restored image data and text data representing a standard image set in advance, every time the restored image data is outputted by said image restoration circuit means.

15. An image processor according to claim 14, further comprising judging means for determining whether a subject image represented by the input image data coincides with at least one of the standard images on the basis of the index values calculated in sequence by said similarity operating circuit means.

16. An image processor according to claim 14, wherein said pseudo-gray-scale image generating circuit means binarizes the converted image data using an error diffusion method and generates dither image data represented by one bit per pixel.

17. An image processor according to claim 14, wherein said pseudo-gray-scale image generating circuit means subjects the converted image data to a four-value conversion using an error diffusion method and generates pseudo-gray-scale image data represented by two bits per pixel.

18. An image processor according to claim 14, wherein

said image restoration circuit means restores the pseudo-gray-scale image data for each and every one of a plurality of small areas, each of which is provided in a window that scans the pseudo-gray-scale image generated by said pseudo-gray-scale image generating circuit means, and outputting restored image data for a window at each and every position;

and

said similarity operating circuit means calculates the index value representing a degree of similarity between the restored image data of each window sequentially outputted by the image restoration circuit means and text data representing a standard image set in advance.

19. An image processor according to claim 18, further comprising peak holding circuit means for detecting an index value, which represents the highest degree of similarity, from among similarity index values regarding each of the windows obtained from said similarity operating circuit means, and position data indicative of the window in which this index value is generated in every predetermined period of time.

20. An image processor according to claim 19, further comprising judging means for determining whether a subject image represented by the input image data has an image portion which coincides with at least one of the standard images on the basis of the index values detected by said peak holding circuit means.

21. An image processor according to claim 18, wherein the index value representing said degree of similarity is a difference value in lightness of the image, and said similarity operating circuit means calculates an absolute value of a difference between restored image data and corresponding text data for each and every small area, sums the absolute values of these differences over all small areas contained in one window and outputs the sum value as a difference value of lightness.

22. An image processor according to claim 21, further comprising a peak holding circuit means for detecting a smallest difference value from among difference values of lightness regarding each of the windows obtained by said similarity operating circuit means, as well as position data indicative of the window in which the smallest difference value is produced.

23. An image processor according to claim 18, wherein said similarity operating circuit means calculates a cross-correlation value between restored image data and text data for each and every window, and outputs this cross-correlation value as the index value representing degree of similarity.

24. An image processor according to claim 23, further comprising a peak holding circuit means for detecting a largest cross-correlation value from among the cross-correlation values regarding each of the windows obtained by said similarity operating circuit means, as well as position data indicative of the window in which the largest cross-correlation value is produced.

25. An image processor comprising:

converting means for converting input image data into image data representing images different in rotational angle or in size from the image represented by the input image data in sequence at a predetermined period of time;

pseudo-gray-scale image generating circuit means for converting the image data converted by said converting means and represented by a first prescribed number of bits per pixel, every time the converted image data is supplied from said converting means, into pseudo-gray-scale image data represented by a second prescribed number of bits, which is less than said first prescribed number of bits;

image restoration circuit means for restoring the pseudo-gray-scale image data for each and every one of a plurality of small areas, each of which comprises a plurality of pixels, with said each small area serving as one restored pixel, every time the pseudo-gray-scale image data is supplied from said pseudo-gray-scale image generating circuit means and outputting restored image data; and

a plurality of similarity operating circuit means, which operate in parallel, each for calculating an index value representing a degree of similarity between the restored image data and text data representing a standard image set in advance, every time the restored image data is outputted by said image restoration circuit means, the standard images for the plurality of similarity operating circuit means being formed based on a single model and different in rotational angle or in size from one another.

26. An image processor accoring to claim 25, further comprising judging means for determining whether a subject image represented by the input image data coincides with at least one of the standard images on the basis of the index values calculated in sequence by the plurality of said similarity operating circuit means.

27. An image processor according to claim 25, wherein said pseudo-gray-scale image generating circuit means binarizes the converted image data using an error diffusion method and generates dither image data represented by one bit per pixel.

28. An image processor according to claim 25, wherein said pseudo-gray-scale image generating circuit means subjects the converted image data to a four-value conversion using an error diffusion method and generates pseudo-gray-scale image data represented by two bits per pixel.

29. An image processor according to claim 25, wherein

said image restoration circuit means restores the pseudo-gray-scale image data for each and every one of a plurality of small areas, each of which is provided in a window that scans the pseudo-gray-scale image generated by said pseudo-gray-scale image generating circuits means, and outputs restored image data for a window at each and every position;

and

said similarity operating circuit means calculates the index value representing a degree of similarity between the restored image data of each window sequentially outputted by the image restoration circuit means and text data representing a standard image set in advance.

30. An image processor according to claim 29, further comprising a plurality of peak holding circuit means each for detecting an index value, which represents the highest degree of similarity, from among similarity index values regarding each of the windows obtained from said corresponding similarity operating circuit means, and position data indicative of the window in which this index value is generated in every predetermined period of time.

31. An image processor according to claim 30, further comprising judging means for determining whether a subject image represented by the input image data has an image portion which coincides with at least one of the standard images on the basis of the index values detected by the plurality of said peak holding circuit means.

32. An image processor according to claim 29, wherein the index value representing said degree of similarity is a difference value in lightness of the image, and said similarity operating circuit means calculates an absolute value of a difference between restored image data and corresponding text data for each and every small area, sums the absolute values of these differences over all small areas contained in one window, and outputs the sum value as a difference value of lightness.

33. An image processor according to claim 32, further comprising a peak holding circuit means for detecting a smallest difference value from among difference values of lightness regarding each of the windows obtained by said similarity operating circuit means, as well as position data indicative of the window in which the smallest difference value is produced.

34. An image processor according to claim 29, wherein said similarity operating circuit means calculates a cross-correlation value between restored image data and text data for each and every window, and outputs this cross-correlation value as the index value representing degree of similarity.

35. An image processor according to claim 34, further comprising a peak holding circuit means for detecting a largest cross-correlation value from among the cross-correlation values regarding each of the windows obtained by said similarity operating circuit means, as well as position data indicative of the window in which the largest cross-correlation value is produced.

36. An image processing method comprising the steps of:

converting input image data of one frame, which is represented by a first prescribed number of bits per pixel, into pseudo-gray-scale image data of one frame represented by a second prescribed number of bits, which is less than the first prescribed number of bits;

summing the pseudo-gray-scale image data for each and every one of a plurality of small areas, each of which comprises a plurality of pixels and is provided in a window that scans the pseudo-gray-scale image generated, thereby restoring the image data for a window at each and every position, with said each small area serving as one restored pixel; and

calculating index values with use of text data of a plurality of standard images which are different from one another for each restored image data of the plurality of frames in parallel, the index value representing a degree of similarity between the restored image data of each window and text data representing the standard image set in advance.

37. An image processing method according to claim 36, further comprising the step of judging whether a subject image represented by the input image data coincides with at least one of the standard images on the basis of the index values calculated.

38. An image processing method comprising the steps of:

converting input image data of one frame into image data representing images of a plurality of frames different in rotational angle or in size from the image represented by the input image data in sequence at a predetermined period of time;

converting the image data of one frame converted by above converting step and represented by a first prescribed number of bits per pixel, every time the converted image data of one frame is supplied, into pseudo-gray-scale image data of one frame represented by a second prescribed number of bits, which is less than the first prescribed number of bits;

summing said pseudo-gray-scale image data of one frame for each and every one of a plurality of small areas, each of which comprises a plurality of pixels and is provided in a window that scans the pseudo-gray-scale image generated, thereby restoring the image data for a window at each and every position, with said each small area serving as one restored pixel, every time the pseudo-gray-scale image data of one frame is supplied; and

calculating an index value representing a degree of similarity between the restored image data of each window and text data representing a standard image set in advance every time the restored image data of one frame is supplied.

39. An image processing method according to claim 38, further comprising the step of judging whether a subject image represented by the input image data coincides with at least one of the standard images on the basis of the index values calculated and supplied in sequence.

40. An image processing method comprising the steps of:

converting input image data of one frame into image data representing images of a plurality of frames different in rotational angle or in size from the image represented by the input image data in sequence at a predetermined period of time;

converting the image data of one frame converted by the above converting step and represented by a first prescribed number of bits per pixel, every time the converted image data of one frame is supplied, into pseudo-gray-scale image data of one frame represented by a second prescribed number of bits, which is less than the first prescribed number of bits;

summing said pseudo-gray-scale image data of one frame for each and every one of a plurality of small areas, each of which comprises a plurality of pixels and is provided in a window that scans the pseudo-gray-scale image generated, thereby restoring the image data for a window at each and every position, with said each small area serving as one restored pixel, every time the pseudo-gray-scale image data of one frame is supplied; and

calculating index values with use of text data of a plurality of standard images which are formed based on a single model and different in rotational angle or in size from one another for each restored image data of the plurality of frames in parallel and every time the restored image data is supplied, the index value representing a degree of similarity between the restored image data of each window and text data representing the standard image set in advance.

41. An image processing method according to claim 40, further comprising the step of judging whether a subject image represented by the input image data coincides with at least one of the standard images on the basis of the index values and supplied in parallel and every time the index value is calculated.