Method, system and computer readable medium for identifying chest radiographs using image mapping and template matching techniques

• US 20020021829A1
• Filed: 03/26/2001
• Published: 02/21/2002
• Est. Priority Date: 03/28/2000
• Status: Active Grant

• Current Assignee: ARCH Development Corporation (University of Chicago)
• Original Assignee: ARCH Development Corporation (University of Chicago)
• Inventors: Morishita, Junji, Arimura, Hidetaka, Doi, Kunio, Katsuragawa, Shigehiko
• Granted Patent: US 6,836,558 B2
• Time in Patent Office: Days
• US Class Current: 382/131
• CPC Class Codes: G06T 2207/30061 Lung
• CPC Class Codes: G06T 7/0002 Inspection of images, e.g. flaw detection
• CPC Class Codes: G06T 7/001 using an image reference approach
• CPC Class Codes: G06T 7/32 using correlation-based methods
• CPC Class Codes: G06V 10/7515 Shifting the patterns to accommodate for positional errors

Abstract

A method, system and computer readable medium for a computer-automated method for identifying given image data, including obtaining template image data corresponding to said given image data; calculating correlation values between the given image data and said template image data; and identifying said image data based on the correlation values calculated in the calculating step.

Claims

1. A computer-automated method for identifying given image data, comprising:

obtaining template image data corresponding to said given image data;

calculating correlation values between the given image data and said template image data; and

identifying said image data based on the correlation values calculated in the calculating step.

2. The method of claim 1, wherein said obtaining step comprises:

obtaining temporal image data purported to be derived from the same patient as the given image data but at an earlier time.

3. The method of claim 2, wherein said obtaining step comprises:

reducing a matrix size of the obtained temporal image data to produce a reduced size matrix image and using the reduced size matrix image as said template image data.

4. The method of claim 1, wherein:

said calculating step comprises,

a) calculating a correlation value based on said given image data and said template image data;

b) shifting one of the given image data and the template image data horizontally and vertically incrementally and for each incremental shift calculating a correlation value therebetween,

c) determining a first best match between the given image data and the template image data based on the correlation values calculated in steps a) and b),

d) rotating incrementally one of the given image data and the template image data yielding the first best match and calculating a correlation value therebetween at each rotational increment, and

e) determining a second best match between the give image data and the template image data based on the correlation values calculated in step d); and

said identifying step comprises identifying said image data based on the correlation value based on the correlation value corresponding to said second best match.

5. The method of claim 2, wherein:

said calculating step comprises,

a) calculating a correlation value based on said given image data and said template image data;

b) shifting one of the given image data and the template image data horizontally and vertically incrementally and for each incremental shift calculating a correlation value therebetween,

c) determining a first best match between the given image data and the template image data based on the correlation values calculated in steps a) and b),

d) rotating incrementally one of the given image data and the template image data yielding the first best match and calculating a correlation value therebetween at each rotational increment, and

e) determining a second best match between the give image data and the template image data based on the correlation values calculated in step d); and

said identifying step comprises identifying said image data based on the correlation value based on the correlation value corresponding to said second best match.

6. The method of claim 3, wherein:

said calculating step comprises,

a) calculating a correlation value based on said given image data and said template image data;

b) shifting one of the given image data and the template image data horizontally and vertically incrementally and for each incremental shift calculating a correlation value therebetween,

c) determining a first best match between the given image data and the template image data based on the correlation values calculated in steps a) and b),

d) rotating incrementally one of the given image data and the template image data yielding the first best match and calculating a correlation value therebetween at each rotational increment, and

e) determining a second best match between the give image data and the template image data based on the correlation values calculated in step d); and

said identifying step comprises identifying said image data based on the correlation value based on the correlation value corresponding to said second best match.

7. The method of claims 1, 2, 3, 4, 5 or 6, wherein said identifying step comprises:

identifying whether the given image data was derived from the same patient from which the template image data was derived.

8. The method of claims 1, 2, 3, 4, 5 or 6, wherein said identifying step comprises:

identifying whether the given image data was derived from the same anatomical structure as that from which the template image data was derived.

9. The method of claim 1, wherein:

said step of obtaining template image data comprises obtaining first and second template image data corresponding to first and second views of a same anatomical structure;

said calculating step comprises calculating correlation values between said image data and said first and second template image data; and

said identifying step comprises identifying said given image data as corresponding to one of said first and second views based on the correlation values obtained in said calculating step.

10. The method of claim 9, comprising:

obtaining original image data of a patient; and

reducing a matrix size of the original image data to produce said given image data.

11. The method of claim 1, wherein said step of obtaining comprises:

obtaining template image data corresponding to averaged image data derived from plural images.

12. The method of claim 9, wherein said step of obtaining comprises:

obtaining first and second template image data each corresponding to averaged image data derived from plural images.

13. The method of claim 9, wherein:

said step of obtaining template image data comprises obtaining first and second template image data corresponding to PA and lateral views, respectively;

said calculating step comprises calculating correlation values between said first and second template image data and said given image data; and

said identifying step comprises identifying whether said given image data corresponds to a PA view or a lateral view in consideration of the correlation values calculated in said calculating step.

14. The method of claim 9, wherein:

said step of obtaining plural template image data comprises obtaining first and second template image data corresponding to PA views and lateral views, respectively, of average sized patients, small-size patients and large-size patients;

said step of calculating correlation values comprises calculating first and second correlation values between said given image data and the obtained first and second template image data; and

said identifying step comprises identifying whether said given image data corresponds to a PA view or a lateral view in consideration of the first and second correlation values.

15. The method of claim 14, wherein said identifying step comprises:

comparing said first and second correlation values with a first threshold and with each other; and

determining, when one of the correlation values exceeds said first threshold and is greater than the other correlation value by a first predetermined amount, that the given image data corresponds to the view of whichever of said first and second template image data yielded said one of said correlation values.

16. The method of claim 15, wherein when one of said first and second correlation values exceeds the first threshold but does not have a value greater than the other correlation value by said first predetermined amount, the method further comprises:

obtaining third and fourth template image data corresponding to corresponding to PA and lateral views of small sized patients, respectively, and fifth and sixth template image data corresponding to PA and lateral views of large sized patients, respectively;

calculating third, fourth, fifth and sixth correlation values between said given image data and said third, fourth, fifth and sixth template image data, respectively; and

determining which of the third through sixth correlation values is the greatest, and identifying the given image data as corresponding to the view of the template image data yielding the greatest correlation value.

17. The method of claim 15, wherein when neither of said first and second correlation values exceeds the first threshold, the method further comprises:

eliminating edge pixels from said given image data and said first and second template image data to produce reduced given image data and reduced first and second template image data;

calculating third and fourth correlation values between said reduced given image data and said reduced first and second template image data; and

said identifying step comprises, comparing said third and fourth correlation values with a second threshold and with each other and

determining, when one of the third and fourth correlation values exceeds said second threshold and is greater than the other of said third and fourth correlation valuess by a second predetermined amount, that the given image data corresponds to the view of whichever of said reduced first and second template image data yielded said one of said third and fourth correlation values.

18. The method of claim 17, wherein when one of said third and fourth correlation values exceeds the second threshold but does not have a value greater than the other of said third and fourth correlation values by the second predetermined amount, the method further comprises:

obtaining third and fourth template image data corresponding to corresponding to PA and lateral views of small sized patients, respectively, and fifth and sixth template image data corresponding to PA and lateral views of large sized patients, respectively;

eliminating edge pixels from said third through sixth template image data to obtain reduced third through sixth template image data;

calculating fifth, sixth, seventh and eighth correlation values between said reduced given image data and said reduced third, fourth, fifth and sixth template image data, respectively; and

determining which of the fifth through eighth correlation values is the greatest, and

identifying the given image data as corresponding to the view of the reduced third through fourth template image data yielding the greatest of said fifth through eight correlation values.

19. The method of claim 14, comprising:

obtaining original image data of a patient; and

reducing a matrix size of the original image data to produce said given image data.

20. The method of claim 14, wherein said step of obtaining comprises:

obtaining first and second template image data each corresponding to averaged image data derived from plural images.

21. The method of claim 15, wherein said step of obtaining comprises:

obtaining first and second template image data each corresponding to averaged image data derived from plural images.

22. The method of claim 9, wherein:

said step of obtaining plural template image data comprises obtaining first template image data corresponding to first and second views of a skull.

23. The method of claim 9, wherein:

said step of obtaining plural template image data comprises obtaining first template image data corresponding to first and second views of a hand.

24. The method of claim 9, wherein:

said step of obtaining plural template image data comprises obtaining first template image data corresponding to first and second views of a foot.

25. The method of claim 9, wherein:

said step of obtaining plural template image data comprises obtaining first template image data corresponding to first and second views of an abdomen.

26. The method of claim 9, wherein:

said step of obtaining plural template image data comprises obtaining first template image data corresponding to first and second ultrasound data.

27. The method of claim 9, wherein:

said step of obtaining plural template image data comprises obtaining first template image data corresponding to first and second computed tomography (CT) data.

28. The method of claim 9, wherein:

said step of obtaining plural template image data comprises obtaining first template image data corresponding to first and second magnetic resonance imaging (MRI) data.

29. An image processing system configured to perform the steps recited in any one of claims 1-6 and 9-28.

30. An image processing system configured to perform the steps recited in claim 7.

31. An image processing system configured to perform the steps recited in claim 8.

32. A storage medium storing a program for performing the steps recited in any one of claims 1-6 and 9-28.

33. A storage medium storing a program for performing the steps recited in claim 7.

34. A storage medium storing a program for performing the steps recited in claim 8.