System and method for detection of lesions in three-dimensional digital medical image

• US 9,014,447 B2
• Filed: 02/26/2010
• Issued: 04/21/2015
• Est. Priority Date: 02/27/2009
• Status: Expired due to Fees

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: Slabaugh, Gregory Gibran, Siddique, Muhammad Musib, Boyes, Richard Graham, Douiri, Abdel, Ye, Xujiong
• Primary Examiner(s): Thirugnanam, Gandhi
• Application Number:
  US12/714,233
• Publication Number:
  US 20100220913A1
• Time in Patent Office: 1,880 Days
• Field of Search: 382/131
• US Class Current: 382/131
• CPC Class Codes: G06T 2207/10072 Tomographic images
• CPC Class Codes: G06T 2207/10132 Ultrasound image
• CPC Class Codes: G06T 2207/20084 Artificial neural networks [ANN]
• CPC Class Codes: G06T 2207/30032 Colon polyp
• CPC Class Codes: G06T 7/0012 Biomedical image inspection

Abstract

A system and method for characterizing a point in a three-dimensional digital medical image includes determining third order derivatives of the three-dimensional digital medical image at one or more points in the three-dimensional digital medical image. At each of the one or more points, one or more invariant or semi-invariant features that characterize a local geometry at the point is determined using the third order derivatives of the three-dimensional digital medical image. The invariant or semi-invariant features are used for automatic detection of lesions, abnormalities, and other anatomical structures of interest in the three-dimensional digital medical image.

Claims

1. A method for characterizing a point in a three-dimensional digital medical image, the method comprising:

determining, by a processor, third order derivatives of the three-dimensional digital medical image at one or more points within the three-dimensional digital medical image;

determining, by the processor, at each of the one or more points, using the third order derivatives of the three-dimensional digital medical image, one or more invariant features that characterize a local geometry at the point, each of the one or more invariant features being invariant with respect to a parameterization of a surface which relates to the local geometry; and

characterizing each of the one or more points using the one or more determined invariant features; and

distinguishing a first area within the three-dimensional medical image which relates to an abnormal tissue from a second area within the three-dimensional medical image which relates to a normal body tissue, based on a contrast therebetween with respect to greyscale representations of points included in the first area and points included in the second area,

wherein the characterizing comprises assigning a respective greyscale representation to each point in the three-dimensional medical image based on the determined third-order derivative for the corresponding point.

2. The method of claim 1, wherein said determining one or more invariant features comprises determining covariant derivatives using the third order derivatives of the three-dimensional digital medical image.

3. The method of claim 1, further comprising, for each of the one or more points, determining the local geometry using an implicit iso-surface.

4. The method of claim 1, wherein said determining the one or more invariant features comprises calculating first, second, and third order derivatives of the three-dimensional digital medical image.

5. The method of claim 1, further comprising identifying an anatomical structure in the three-dimensional digital medical image using the one or more invariant features.

6. The method of claim 1, further comprising identifying a lesion in the three-dimensional digital medical image using the one or more invariant features.

7. The method of claim 6, wherein said identifying a lesion in the three-dimensional digital medical image includes using the one or more invariant features to reduce false positives.

8. The method of claim 1, further comprising using the one or more invariant features to segment the three-dimensional digital medical image into a plurality of labels.

9. The method of claim 1, further comprising deriving, using the one or more invariant features, a spatial transformation for aligning the three-dimensional digital medical image with one or more further image.

10. The method of claim 9, further comprising applying the spatial transformation to the three-dimensional digital medical image to align the three-dimensional digital medical image with the one or more further image.

23. The method of claim 1, wherein at least one of the invariant features is derived based on applying at least one of a plurality of Codazzi equations to an expression relating to a regular parameterization of a surface of the three-dimensional digital medical image,

wherein the Codazzi equations yield coefficients for a symmetric covariant 3-tensor quantity.

11. A method for characterizing a point in a three dimensional digital medical image, the method comprising:

determining, by a processor, third order derivatives of the three-dimensional digital medical image, at one or more points within the three-dimensional digital medical image;

determining, by the processor, at each of the one or more points, using the third order derivatives of the three-dimensional digital medical image, one or more semi-invariant features that characterize a local geometry at the point, each of the one or more semi-invariant features being calculated using Cartesian coordinate directions to yield an approximation of a corresponding invariant feature which is invariant with respect to a parameterization of a surface which relates to the local geometry;

characterizing each of the one or more points using the one or more determined semi-invariant features; and

distinguishing a first area within the three-dimensional medical image which relates to an abnormal tissue from a second area within the three-dimensional medical image which relates to a normal body tissue, based on a contrast therebetween with respect to greyscale representations of points included in the first area and points included in the second area,

wherein the characterizing comprises assigning a respective greyscale representation to each point in the three-dimensional medical image based on the determined third-order derivative for the corresponding point.

12. The method of claim 11, wherein said determining one or more semi-invariant features comprises determining derivatives of the three-dimensional digital medical image in Cartesian coordinates.

13. The method of claim 11, further comprising, for each of the one or more points, determining the local geometry using an implicit iso-surface.

14. The method of claim 11, wherein said determining one or more semi-invariant features comprises calculating first, second, and third order derivatives of the three-dimensional digital medical image.

15. The method of claim 11, further comprising identifying an anatomical structure in the three-dimensional digital medical image using the one or more semi-invariant features.

16. The method of claim 11, further comprising identifying a lesion in the three-dimensional digital medical image using the one or more semi-invariant features.

17. The method of claim 16, wherein said identifying a lesion in the three-dimensional digital medical image includes using the one or more semi-invariant features to reduce false positives.

18. The method of claim 11, further comprising using the one or more semi-invariant features to segment the three-dimensional digital medical image into a plurality of labels.

19. The method of claim 11, further comprising deriving, using the one or more semi-invariant features, a spatial transformation for aligning the three-dimensional digital medical image with one or more further image.

20. The method of claim 19, further comprising applying the spatial transformation to the three-dimensional digital medical image to align the three-dimensional digital medical image with the one or more further image.

24. The method of claim 11, wherein at least one of the semi-invariant features is derived based on using Cartesian coordinate directions in conjunction with applying at least one of a plurality of Codazzi equations to an expression relating to a regular parameterization of a surface of the three-dimensional digital medical image,

wherein the Codazzi equations yield coefficients for a symmetric covariant 3-tensor quantity.

21. An article of manufacture including a non-transitory computer readable storage medium having stored thereon computer-executable instructions, execution of which by a computing device causes the computing device to perform processes comprising:

determining third order derivatives of a three-dimensional digital medical image, at one or more points within the three-dimensional digital medical image;

determining, at each of the one or more points, one or more invariant or semi-invariant features that characterize a local geometry at the point using the third order derivatives of the three-dimensional digital medical image, each of the one or more invariant features being invariant with respect to a parameterization of a surface which relates to the local geometry, and each of the one or more semi-invariant features being calculated using Cartesian coordinate directions to yield an approximation of a corresponding one of the one or more invariant features which is invariant with respect to the parameterization of the surface which relates to the local geometry;

characterizing each of the one or more points using the one or more determined invariant features; and

distinguishing a first area within the three-dimensional medical image which relates to an abnormal tissue from a second area within the three-dimensional medical image which relates to a normal body tissue, based on a contrast therebetween with respect to greyscale representations of points included in the first area and points included in the second area,

wherein the characterizing comprises assigning a respective greyscale representation to each point in the three-dimensional medical image based on the determined third-order derivative for the corresponding point.

22. An apparatus comprising:

a communication interface which is configured to receive scan data; and

a processor which is configured to determine third order derivatives of a three-dimensional digital medical image which is based on the received scan data, at one or more points within the three-dimensional digital medical image, and to determine, at each of the one or more points, one or more invariant or semi-invariant features that characterize a local geometry at the point using the third order derivatives of the three-dimensional digital medical image, each of the one or more invariant features being invariant with respect to a parameterization of a surface which relates to the local geometry, and each of the one or more semi-invariant features being calculated using Cartesian coordinate directions to yield an approximation of a corresponding one of the one or more invariant features which is invariant with respect to the parameterization of the surface which relates to the local geometry,

wherein the processor is further configured to distinguish a first area within the three-dimensional medical image which relates to an abnormal tissue from a second area within the three-dimensional medical image which relates to a normal body tissue, based on a contrast therebetween with respect to greyscale representations of points included in the first area and points included in the second area, the greyscale representations being assigned to each point in the three-dimensional medical image based on the determined third-order derivative for the corresponding point.