Method and system for automated detection of microcalcification clusters in mammograms

• US 5,574,799 A
• Filed: 06/12/1992
• Issued: 11/12/1996
• Est. Priority Date: 06/12/1992
• Status: Expired due to Term

• Current Assignee: Johns Hopkins University
• Original Assignee: Johns Hopkins University
• Inventors: Christens-Barry, William A., Bankman, Isaac N.
• Primary Examiner(s): Boudreau, Leo
• Assistant Examiner(s): ANDERSON, DAVID
• Application Number:
  US07/897,650
• Time in Patent Office: 1,614 Days
• Field of Search: 382/6, 382/1, 382/16, 382/22, 382/28, 364/413.01, 364/413.02, 364/413.13, 364/413.14, 364/413.19, 348/77, 348/79
• US Class Current: 382/132
• CPC Class Codes: G06V 10/40 Extraction of image or video features

Abstract

The invention is a novel method for automated detection of objects in images. One specific use is to detect malignant microcalcification clusters in mammograms. The method operates by forming a contour plot of the image (mammogram), the object (microcalcification) in the contour plot being comprised of a set of nested contour lines wherein the contour lines correspond to intensity levels thereby causing the object to appear as a prominent peak in relation to the local surround in the image.

Claims

1. A method of operating an image interpretation system wherein an object in the image is automatically detected, the method comprising the steps of:

digitizing the image;

inputting the digitized image into a data processing means;

detecting automatically the object using the data processing means and the digitized image comprising the steps of;

forming a contour plot of the image, the object in the contour plot comprised of a set of nested contour lines wherein the contour lines correspond to intensity levels thereby causing the object to appear as a peak in relation to the surrounding background in the image;

obtaining a sequence of contour areas from the set of nested contour lines progressing from the contour line having the highest intensity level to the contour line having the lowest intensity level;

computing a feature of the set of nested contour lines using the sequence of contour areas; and

detecting the object using the computed feature; and

indicating the location of the object in the image.

2. The method as recited in claim 1, wherein the object is a microcalcification.

3. The method as recited in claim 2, wherein the image is a mammogram.

4. The method as recited in claim 3, wherein the mammogram is digitized at a spatial resolution of approximately 25 microns per pixel.

5. The method as recited in claim 1, wherein each contour line is formed of pixels having intensity values equal to or higher than the intensity level of the contour line.

6. The method as recited in claim 5, wherein the computing a feature step comprises the step of computing three features of the set of nested contour lines using the sequence of contour areas.

7. The method as recited in claim 6, wherein the three features are departure, prominence and steepness.

8. The method as recited in claim 6, wherein the object is a cluster of a plurality of microcalcifications.

9. The method as recited in claim 8, wherein the detecting the object step comprises the step of selecting the microcalcification cluster when three or more peaks occur in an area of predetermined size and have values of departure, prominence and steepness that fall in the corresponding acceptance range.

10. The method as recited in claim 9, wherein the area of predetermined size is about 1 cm².

11. The method as recited in claim 5, the forming a contour plot step further comprising the step of determining the highest pixel intensity value of the image, the lowest pixel intensity value of the image, and the mean pixel intensity value of the image.

12. The method as recited in claim 11, wherein the spacing of the levels of intensity of the contour lines is equal and is preset.

13. The method as recited in claim 11, wherein the spacing of the levels of intensity of the contour lines is equal and is set adaptively for each image.

14. The method as recited in claim 13, wherein the spacing is in proportion to the range of the intensity levels of the contour lines.

15. The method as recited in claim 14, wherein the contour line having the largest intensity level is set to equal the highest pixel intensity value of the image minus the spacing and the contour line having the lowest intensity level is set to equal the lowest pixel intensity value of the image.

16. The method as recited in claim 15, wherein the contour line having the lowest intensity level is set to equal the mean pixel intensity value of the image.

17. The method as recited in claim 5, wherein a contour line is ignored if the area enclosed by the contour is smaller than a first threshold.

18. The method as recited in claim 17, wherein the first threshold is set to the area of a square about 0.1 mm on a side.

19. The method as recited in claim 5, wherein the sequence stops at a contour line enclosing an area larger than a second threshold.

20. The method as recited in claim 19, wherein the second threshold is set to the area of a square about 2 mm on a side.

21. The method as recited in claim 5, wherein the computing a feature step comprises the step of computing a value for the departure of the peak from the surrounding background in the image by computing the rate of change of the sequence of contour areas as the sequence approaches the contour line having the lowest intensity level.

22. The method as recited in claim 21, wherein the departure value is set to the highest relative increase in the first difference sequence of contour areas and the corresponding intensity level is the departure level.

23. The method as recited in claim 5, wherein the computing a feature step comprises the step of computing a prominence of the peak, the prominence being an integer equal to the number of contour lines above the departure level.

24. The method as recited in claim 5, wherein the computing a feature step comprises the step of computing a steepness of the peak.

25. The method as recited in claim 24, wherein the steepness computing step comprises the steps of:

computing a ratio of change in grey value to change in area with increasing contour lines at each level above the departure level; and

setting the steepness to the mean of the ratios.

26. A method of operating a mammogram interpretation system wherein a microcalcification in the mammogram is automatically detected, the method comprising the steps of:

digitizing the mammogram;

inputting the digitized mammogram into a data processing means;

detecting automatically the microcalcification using the data processing means and the digitized mammogram comprising the steps of;

forming a contour plot of the mammogram, the microcalcification in the contour plot comprised of a set of nested contour lines wherein each contour line is formed of pixels having intensity values equal to or higher than the intensity level of the contour line thereby causing the microcalcification to appear as a prominent peak in relation to the surrounding background in the mammogram;

obtaining a sequence of contour areas from the set of nested contour lines progressing from the contour line having the highest intensity level to the contour line having the lowest intensity level;

computing a value for the departure of the peak from the surrounding background in the mammogram by computing the rate of change of the sequence of contour areas as the sequence approaches the contour line having the lowest intensity level, wherein the departure value is set to the highest relative increase in the first difference sequence of contour areas and the corresponding intensity level is the departure level;

computing the prominence of the peak, the prominence being an integer equal to the number of contour lines above the departure level;

computing the steepness of the peak comprising the steps of;

computing a ratio of change in grey value to change in area with increasing contour lines at each level above the departure level; and

setting the steepness to the mean of the ratios; and

detecting a microcalcification cluster of three microcalcifications when three or more peaks occur in an area of about 1 cm² and have values of departure, prominence and steepness that fall in the corresponding acceptance range; and

indicating the location of the microcalcification cluster in the mammogram.

27. An image interpretation system wherein an object in the image is automatically detected, the system comprising:

means for digitizing the image;

a data processing means for receiving the digitized image;

means for automatically detecting the object using the data processing means and the digitized image, the automatic detection means comprising;

means for forming a contour plot of the image, the object in the contour plot comprised of a set of nested contour lines wherein the contour lines correspond to intensity levels thereby causing the object to appear as a peak in relation to the surrounding background in the image;

means for obtaining a sequence of contour areas from the set of nested contour lines progressing from the contour line having the highest intensity level to the contour line having the lowest intensity level;

means for computing a feature of the set of nested contour lines using the sequence of contour areas; and

means for detecting the object using the computed feature; and

means for indicating the location of the object in the image.

28. The system as recited in claim 27, wherein the object is a microcalcification.

29. The system as recited in claim 28, wherein the image is a mammogram.

30. The system as recited in claim 27, wherein each contour line is formed of pixels having intensity values equal to or higher than the intensity level of the contour line.

31. The method as recited in claim 30, wherein the computing a feature means comprises means for computing three features of the set of nested contour lines using the sequence of contour areas.

32. The system as recited in claim 31, wherein the three features are departure, prominence and steepness.

33. The system as recited in claim 32, wherein the object is a cluster of a plurality of microcalcifications.

34. The system as recited in claim 33, wherein the detecting means comprises means for selecting the microcalcification cluster when three or more peaks occur in an area of about 1 cm² and have values of departure, prominence and steepness that fall in the corresponding acceptance range.

35. An mammogram interpretation system wherein a microcalcification in the mammogram is automatically detected, the system comprising:

means for digitizing the mammogram;

a data processing means for receiving the digitized mammogram;

means for automatically detecting the microcalcification using the data processing means and the digitized mammogram, the automatic detection means comprising;

means for forming a contour plot of the mammogram, the microcalcification in the contour plot comprised of a set of nested contour lines wherein each contour line is formed of pixels having intensity values equal to or higher than the intensity level of the contour line thereby causing the microcalcification to appear as a prominent peak in relation to the surrounding background in the mammogram;

means for obtaining a sequence of contour areas from the set of nested contour lines progressing from the contour line having the highest intensity level to the contour line having the lowest intensity level;

means for computing a value for the departure of the peak from the surrounding background in the mammogram by computing the rate of change of the sequence of contour areas as the sequence approaches the contour line having the lowest intensity level, wherein the departure value is set to the highest relative increase in the first difference sequence of contour areas and the corresponding intensity level in the departure level;

means for computing the prominence of the peak, the prominence being an integer equal to the number of contour lines above the departure level;

means for computing the steepness of the peak comprising;

means for computing a ratio of change in grey value to change in area with increasing contour lines at each level above the departure level; and

means for setting the steepness to the mean of the ratios; and

means for detecting a microcalcification cluster of three microcalcifications when three or more peaks occur in an area of about 1 cm² and have values of departure, prominence and steepness that fall in the corresponding acceptance range; and

means for indicating the location of the microcalcification cluster in the mammogram.