Three-dimensional reconstructions of a breast from two X-ray mammographics

• US 7,519,212 B2
• Filed: 01/31/2001
• Issued: 04/14/2009
• Est. Priority Date: 03/17/2000
• Status: Expired due to Fees

• Current Assignee: Siemens Medical Solutions USA Incorporated (Siemens AG)
• Original Assignee: Siemens Medical Solutions USA Incorporated (Siemens AG)
• Inventors: Highnam, Ralph, Yam, Shuk Wah Margaret, Brady, John Michael
• Primary Examiner(s): Tabatabai; Abolfazl
• Application Number:
  US10/221,778
• Publication Number:
  US 20040094167A1
• Time in Patent Office: 2,995 Days
• Field of Search: 382128-134, 382/154, 378/37, 378/208, 378/92, 378/98.4, 378/98.6, 378/101, 378/140, 128/915, 128/922, 128/916, 600/407, 600/419, 600/425, 600/443, 600/456, 600/410, 600/411, 600/412, 600/607
• US Class Current: 382/132
• CPC Class Codes: G06T 2207/30068 Mammography; Breast
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/55 from multiple images

Abstract

Methods are described for the production of a three-dimensional reconstruction of a undeformed object from two different views of the object under deformation using a volume constraint and also by matching corresponding features in the two images. The volume constraint involves assuming that the deformed volume is the same as the undeformed volume, and calculating the deformed volume from one of the images. Further, the deformation of the object can be parameterised by finding corresponding image entities in the each of the images. The method is particularly applicable to breast mammograms in which case the two images are the cranio-caudal (CC) image and medio-lateral oblique (MLO) image whose angular separation varies from 35 to 60 degrees. The image entities which are detected in the two images are microcalcifications, and these are matched by detecting a value representing their volume a looking for matches in this value between the two images.

Claims

1. A method of producing a three dimensional representation of an undeformed object, the method comprising:

using a computer to perform steps including;

combining information from two 2D images taken from different viewpoints of the object under deformation, estimating the volume of the deformed object, and constraining the three dimensional representation of the object to have substantially the same volume, wherein the information from each of the two images is combined by the steps of;

(a) detecting the outline of the object in each of the two images, (b) reducing area of the outlined areas by a predetermined amount, and (c) using the outlines of the reduced areas as profiles from different directions of the three dimensional representation of the object.

2. A method according to claim 1, wherein the deformation of the object differs between the two images.

3. A method according to claim 1 wherein the volume of the deformed object is estimated from one of the images.

4. A method according to claim 3 wherein the volume of the deformed object is estimated by summing over the image the volume of slices of the object parallel to the imaging direction.

5. A method according to claim 1 wherein the volume is estimated by assuming at least part of the surface of the deformed object to be a parametric surface.

6. A method according to claim 1 further comprising the steps of:

(d) calculating the volume of the three dimensional representation of the object, (e) comparing it to the estimated volume of the deformed object, and iterating steps (b), (c), (d) and (e) until the volume of the three dimensional representation of the object is substantially equal to the estimated volume of the deformed object.

7. A method according to claim 1 wherein the three dimensional representation of the object comprises parametric surfaces passing through the said profiles.

8. A method according to claim 1 wherein the outlines of the reduced areas are used as profiles from orthogonal directions.

9. A method according to claim 1 wherein the amounts of deformation of the object differs between the two images and the predetermined amounts by which the outlined areas are reduced in the two images differ in accordance with the respective amounts of deformation.

10. A method according to claim 1 wherein the object is deformed parallel to one of the imaging directions.

11. A method according to claim 1 wherein the object is deformed by compression.

12. A method according to claim 1 wherein the object is a human breast.

13. A method according to claim 12 wherein the images are breast mammograms.

14. A method according to claim 13 wherein the images are taken in the cranio-caudal (CC) and medio-lateral oblique (MLO) directions or CC and lateral-medial (LM) views.

15. A method according to claim 14 wherein the volume of the deformed breast is estimated from the CC or MLO or LM image.

16. A method according to claim 1 further comprising parameterising the deformation of the object.

17. A computer program embodied in a computer readable medium and comprising program code adapted to perform the method of claim 1.

18. A method of detecting correspondences between microcalcifications in two mammogram images of a breast, the method comprising:

using a computer to perform steps including;

converting the two images into an h_int representation representing the thicknesses of interesting tissue and fat in regions of the breast contributing to the mammograms, calculating a value v_int representing the interesting volume for each microcalcification based on a sum of the values of h_int for all pixels within the image of the microcalcification, comparing the values of v_int for each microcalcification in one image with each microcalcification in the other image to detect as corresponding those whose v_int values match to a predetermined degree.

19. A method according to claim 18 wherein the calculation of the value v_int for each microcalcification comprises summing for all pixels within the image of the microcalcification the value of h_int multiplied by the area of the pixel.

20. A method according to claim 18 wherein the calculation of the value v_int for each microcalcification further comprises deducting the contribution of non-calcified tissue within the area of the image of the microcalcification.

21. A method according to claim 20 wherein the contribution of non-calcified tissue is estimated on the basis of the value of h_int in the area of the image surrounding the microcalcification.

22. A method according to claim 20 wherein the contribution of non-calcified tissue is estimated on the basis of the average of the value of h_int in the area of the image surrounding the microcalcification.

23. A method according to claim 20 wherein the contribution by non-calcified tissue is calculated by converting the value of h_int in the area of the image surrounding the microcalcification into a volume representative value by multiplying the values of h_int for each pixel in the surrounding area by the area of each pixel and the number of pixels in the surrounding area.

24. A computer program embodied in a computer-readable medium and comprising program code adapted to perform the method of claim 18.