Computer-assisted karyotyping

• US 9,336,430 B2
• Filed: 06/19/2013
• Issued: 05/10/2016
• Est. Priority Date: 06/19/2012
• Status: Active Grant

• Current Assignee: Health Discovery Corporation
• Original Assignee: Health Discovery Corporation
• Inventors: Zhang, Hong, Albitar, Maher
• Primary Examiner(s): SHAH, UTPAL D
• Application Number:
  US13/922,184
• Publication Number:
  US 20140016843A1
• Time in Patent Office: 1,056 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06V 20/69 Microscopic objects, e.g. biological cells or cellular parts
• CPC Class Codes: G06V 20/695 Preprocessing, e.g. image segmentation
• CPC Class Codes: G06V 20/698 Matching; Classification

Abstract

A system and method for computer-assisted karyotyping includes a processor which receives a digitized image of metaphase chromosomes for processing in an image processing module and a classifier module. The image processing module may include a segmenting function for extracting individual chromosome images, a bend correcting function for straightening images of chromosomes that are bent or curved and a feature selection function for distinguishing between chromosome bands. The classifier module, which may be one or more trained kernel-based learning machines, receives the processed image and generates a classification of the image as normal or abnormal.

Claims

1. A method for computer-assisted karyotyping, comprising:

inputting a digitized image of metaphase chromosomes into a processor comprising an image processing module and a classifier module, wherein the image processing module comprises;

a segmenting function adapted for segmenting the digitized image for extracting individual chromosome images, wherein the segmenting function comprises measuring a contour of a chromosome border to detect extremes corresponding to telomeres and centromeres;

a bend correcting function adapted to straighten images of chromosomes that are bent or curved, wherein the bend correction function comprises an algorithm adapted to conformally map an ideal chromosome to the individual chromosome image;

a feature selection function adapted for distinguishing between chromosome bands;

wherein the classifier module generates a classification of the image as normal or abnormal and generates an output therefrom.

2. The method of claim 1, wherein the classifier module comprises at least one kernel-based learning machine.

3. The method of claim 2, wherein the classifier module comprises a plurality of kernel-based learning machines, wherein each kernel-based learning machine classifies a different group of chromosomes.

4. The method of claim 3, wherein the classifier module further comprises a combined kernel-based learning machine for receiving an output of each of the plurality of kernel-based learning machines to generate a combined result for all chromosomes.

5. The method of claim 2, wherein the kernel-based learning machine is a support vector machine.

6. The method of claim 1, wherein the image processing module further comprises:

a medial axis transform function for generating a topological skeleton corresponding to a chromosome image.

7. The method of claim 1, wherein the feature selection function comprises one or more functions selected from the group consisting of Fourier analysis, wavelet transform and 3D distance function.

23. The method of claim 1, wherein the bend correcting function comprises polynomial least squares fitting.

8. A system for computer-assisted karyotyping, comprising:

a data storage device for receiving and storing digitized images of metaphase chromosomes;

a processor in communication with the data storage device comprising an image processing module and a classifier module, wherein the image processing module comprises;

a segmenting function adapted for segmenting the digitized image for extracting individual chromosome images, wherein the segmenting function comprises measuring a contour of a chromosome border to detect extremes corresponding to telomeres and centromeres;

a bend correcting function adapted to straighten images of chromosomes that are bent or curved, wherein the bend correction function comprises an algorithm adapted to conformally map an ideal chromosome to the individual chromosome image;

a feature selection function adapted for distinguishing between chromosome bands;

an output device in communication with the processor for displaying an output comprising a classification of the image as normal or abnormal.

9. The system of claim 8, wherein the classifier module comprises at least one kernel-based learning machine.

10. The system of claim 9, wherein the kernel-based learning machine is a support vector machine.

11. The system of claim 9, wherein the classifier module comprises a plurality of kernel-based learning machines, wherein each kernel-based learning machine classifies a different group of chromosomes.

12. The system of claim 9, wherein the classifier module further comprises a combined support vector machine for receiving an output of each of the plurality of support vector machine to generate a combined result for all chromosomes.

13. The system of claim 8, wherein the image processing module further comprises:

a medial axis transform function for generating a topological skeleton corresponding to a chromosome image.

14. The system of claim 8, wherein the feature selection function comprises one or more functions selected from the group consisting of Fourier analysis, wavelet transform and 3D distance function.

15. The system of claim 8, further comprising storing the output in the data storage device.

19. The non-transitory machine-readable medium of claim 15, wherein the classifier module comprises a plurality of kernel-based learning machines, wherein each kernel-based learning machine classifies a different group of chromosomes.

24. The system of claim 8, wherein the bend correcting function comprises polynomial least squares fitting.

16. A non-transitory machine-readable medium comprising a plurality of instructions that, in response to being executed, result in a computing system executing an automated karyotyping process, comprising:

receiving a digitized image of metaphase chromosomes into a processor comprising an image processing module and a classifier module, wherein the image processing module comprises;

a segmenting function adapted for segmenting the digitized image for extracting individual chromosome images, wherein the segmenting function comprises measuring a contour of a chromosome border to detect extremes corresponding to telomeres and centromeres;

a bend correcting function adapted to straighten images of chromosomes that are bent or curved, wherein the bend correction function comprises an algorithm adapted to conformally map an ideal chromosome to the individual chromosome image;

a feature selection function adapted for distinguishing between chromosome bands;

classifying the processed image from the image processing module to separate normal and abnormal; and

outputting a report to a display device, the report comprising a classification of the image as normal or abnormal.

17. The non-transitory machine-readable medium of claim 16, wherein the classifier module comprises at least one kernel-based learning machine.

18. The non-transitory machine-readable medium of claim 17, wherein the kernel-based learning machine is a support vector machine.

20. The non-transitory machine-readable medium of claim 16, wherein the classifier module further comprises a combined kernel-based learning machine for receiving an output of each of the plurality of kernel-based learning machines to generate a combined result for all chromosomes.

21. The non-transitory machine-readable medium of claim 16, wherein the image processing module further comprises:

a medial axis transform function for generating a topological skeleton corresponding to a chromosome image.

22. The non-transitory machine-readable medium of claim 16, wherein the feature selection function comprises one or more functions selected from the group consisting of Fourier analysis, wavelet transform and 3D distance function.

25. The non-transitory machine-readable medium of claim 16, wherein the bend correcting function comprises polynomial least squares fitting.