Method and system for the segmentation and classification of lesions

• US 6,138,045 A
• Filed: 08/07/1998
• Issued: 10/24/2000
• Est. Priority Date: 08/07/1998
• Status: Expired due to Term

• Current Assignee: ARCH Development Corporation (University of Chicago)
• Original Assignee: ARCH Development Corporation (University of Chicago)
• Inventors: Giger, Maryellen L., Kupinski, Matthew A.
• Primary Examiner(s): Casler, Brian L.
• Application Number:
  US09/131,162
• Time in Patent Office: 809 Days
• Field of Search: 600/407, 600/410, 600/425, 600/437, 128/920, 128/922, 128/915, 382/128, 382/131, 382/154, 382/168-172, 382/190-199, 345/418, 345/429, 345/419, 378/37, 378/62
• US Class Current: 600/425
• CPC Class Codes: G06T 2207/10116 X-ray image
• CPC Class Codes: G06T 2207/20156 Automatic seed setting
• CPC Class Codes: G06T 2207/30068 Mammography; Breast
• CPC Class Codes: G06T 2207/30101 Blood vessel; Artery; Vein; Vascular
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/11 Region-based segmentation
• CPC Class Codes: G06T 7/12 Edge-based segmentation
• CPC Class Codes: G06T 7/143 involving probabilistic approaches, e.g. Markov random field [MRF] modelling

Abstract

A method for the automated segmentation of an abnormality in a medical image, including acquiring first image data representative of the medical image; locating a suspicious site at which the abnormality may exist; establishing a seed point within the suspicious site; and preprocessing the suspicious site with a constraint function to produce second image data in which pixel values distant of the seed point are suppressed. Preprocessing includes using an isotropic Gaussian function centered on the seed point as the constraint function, or for example using an isotropic three dimensional Gaussian function centered on the seed point as the constraint function. The method further includes applying plural thresholds to the second image data to partition the second image data at each threshold; identifying corresponding first image data for the partitioned second image data obtained at each respective threshold; determining a respective index for each of the partitioned first image data; and determining a preferred partitioning, for example that partitioning leading to a maximum index value, based on the indices determined at each threshold, and segmenting the lesion based on the partitioning established by the threshold resulting in the maximum index. If desired, the first image data with the partitioning defined by the threshold which is determined to result in the maximum index, is then displayed. A system and computer readable storage medium are also provided, likewise using the radial gradient index (RGI) or a simple probabilistic models to segment mass lesions, or other similar nodular structures, from surrounding background. In the system, a series of image partitions is likewise created using gray-level information as well as prior knowledge of the shape of typical mass lesions. When the RGI is used, the partition that maximizes RGI is selected. When a probability model is used, probability distributions for gray-levels inside and outside the partitions are estimated, and subsequently used to determine the probability that the image occurred for each given partition. The partition that maximizes this probability is selected as the final lesion partition (contour).

Claims

1. In a method for the automated segmentation of an abnormality in a medical image, the improvement comprising:

acquiring first image data representative of the medical image;

locating a suspicious site at which the abnormality may exist;

establishing a seed point within the suspicious site;

applying a constraint function relative to the suspicious site based on the location of the seed point; and

preprocessing the suspicious site with the constraint function, including multiplying the first image by the constraint function, to produce second image data in which pixel values distant of the seed point are suppressed.

2. The method of claim 1, wherein the preprocessing step comprises:

preprocessing the suspicious site with an isotropic Gaussian function centered on the seed point as the constraint function to produce the second image data.

3. The method of claim 1, wherein the preprocessing step comprises:

preprocessing the suspicious site with an isotropic three dimensional Gaussian function centered on the seed point as the constraint function to produce the second image data.

4. The method according to claim 1, further comprising:

partitioning the site using the second image data to identify pixels in the first image data; and

calculating an index based on the identified pixels in the first image data.

5. The method according to claim 4, comprising:

the partitioning step comprising applying plural thresholds to the second image data to partition the second image data at each threshold;

identifying corresponding first image data for the partitioned second image data obtained at each respective threshold; and

determining a respective index for each of the partitioned first image data; and

determining a preferred partitioning based on the indices determined at each threshold in the preceding step.

6. The method of claim 5, wherein the step of determining a preferred partitioning comprises:

determining which of the determined indices has a maximum value; and

segmenting the lesion based on the partitioning established by the threshold resulting in the maximum index.

7. The method of claim 5, comprising:

displaying the first image data with the partitioning defined by the threshold which is determined to result in the maximum index.

8. The method of any one of claims 4-7, wherein the step of calculating the index comprises:

calculating a radial gradient index.

9. The method of any one of claims 4-7, wherein the step of calculating the index comprises:

calculating the index with a probabilistic measure.

10. The method of any one of claim 5-7, wherein the step of calculating the index comprises:

calculating contributions to the index with a probabilistic measure, including computing contributions to the index based on pixel values both within and outside of a partitioned region established at respective of the thresholds.

11. In a system for the automated segmentation of an abnormality in a medical image, the improvement comprising:

a memory having embodied therein medical image information; and

a processor coupled to the memory the processor configured to;

acquire first image data representative of the medical image;

locate a suspicious site at which the abnormality may exist;

apply a constraint function relative to the suspicious site based on the location of a seed point; and

preprocess the suspicious site with the constraint function and multiply the first image by the constraint function, to produce second image data in which pixel values distant of the seed point are suppressed.

12. The system of claim 11, wherein the processor is further configured to:

preprocess the suspicious site with an isotropic Gaussian function centered on the seed point as the constraint function to produce the second image data.

13. The system of claim 11, wherein the processor is further configured to:

preprocess the suspicious site with an isotropic three dimensional Gaussian function centered on the seed point as the constraint function to produce the second image data.

14. The system according to claim 11, wherein the processor is further configured to:

partition the site using the second image data to identify pixels in the first image data; and

calculate an index based on the identified pixels in the first image data.

15. The system according to claim 14, wherein the processor is further configured to:

apply plural thresholds to the second image data to partition the second image data at each threshold;

identify corresponding first image data for the partitioned second image data obtained at each respective threshold;

determine a respective index for each of the partitioned first image data; and

determine a preferred partitioning based on the indices determined at each threshold.

16. The system of claim 15, wherein the processor is further configured to:

determine which of the determined indices has a maximum value; and

segment the lesion based on the partitioning established by the threshold resulting in the maximum index.

17. The system of claim 15, comprising:

displaying the first image data with the partitioning defined by the threshold which is determined to result in the maximum index.

18. The system of any one of claims 14-17, wherein the processor is further configured to:

calculate a radial gradient index.

19. The system of any one of claims 14-17, wherein the processor is further configured to:

calculate the index with a probabilistic measure.

20. The system of any one of claim 15-17, wherein the processor is further configured to:

calculate contributions to the index with a probabilistic measure and to compute contributions to the index based on pixel values both within and outside of a partitioned region established at respective of the thresholds.

21. A computer readable medium storing computer instructions for the automated segmentation of an abnormality in a medical image, by performing the steps of:

acquiring first image data representative of the medical image;

locating a suspicious site at which the abnormality may exist;

establishing a seed point within the suspicious site;

applying a constraint function relative to the suspicious site based on the location of the seed point; and

preprocessing the suspicious site with the constraint function, including multiplying the first image by the constraint function, to produce second image data in which pixel values distant of the seed point are suppressed.

22. The computer readable medium of claim 21, wherein the preprocessing step comprises:

preprocessing the suspicious site with an isotropic Gaussian function centered on the seed point as the constraint function to produce the second image data.

23. The computer readable medium of claim 21, wherein the preprocessing step comprises:

preprocessing the suspicious site with an isotropic three dimensional Gaussian function centered on the seed point as the constraint function to produce the second image data.

24. The computer readable medium according to claim 21, storing further instructions for performing the steps comprising:

partitioning the site using the second image data to identify pixels in the first image data; and

calculating an index based on the identified pixels in the first image data.

25. The computer readable medium according to claim 24, storing further instructions for performing the steps comprising:

in the partitioning step, applying plural thresholds to the second image data to partition the second image data at each threshold;

identifying corresponding first image data for the partitioned second image data obtained at each respective threshold; and

determining a respective index for each of the partitioned first image data; and

determining a preferred partitioning based on the indices determined at each threshold in the preceding step.

26. The computer readable medium of claim 25, wherein the step of determining a preferred partitioning comprises:

determining which of the determined indices has a maximum value; and

segmenting the lesion based on the partitioning established by the threshold resulting in the maximum index.

27. The computer readable medium of claim 25, storing further instructions for performing the steps comprising:

displaying the first image data with the partitioning defined by the threshold which is determined to result in the maximum index.

28. The computer readable medium of any one of claims 24-27, wherein the step of calculating the index comprises:

calculating a radial gradient index.

29. The computer readable medium of any one of claims 24-27, wherein the step of calculating an index comprises:

calculating the index with a probabilistic measure.

30. The computer readable medium of any one of claim 25-27, wherein the step of calculating an index comprises:

calculating contributions to the index with a probabilistic measure, including computing contributions to the index based on pixel values both within and outside of a partitioned region established at respective of the thresholds.