Systems and methods for computer aided detection of spinal curvature using images and angle measurements

• US 20070055178A1
• Filed: 08/21/2006
• Published: 03/08/2007
• Est. Priority Date: 09/07/2005
• Status: Active Grant

• Current Assignee: Siemens Medical Solutions USA Incorporated (Siemens AG)
• Original Assignee: Siemens Medical Solutions USA Incorporated (Siemens AG)
• Inventors: Odry, Benjamin, Verre, Jeanne, Novak, Carol, Reisman, James
• Granted Patent: US 7,715,605 B2
• Time in Patent Office: Days
• US Class Current: 600/594
• CPC Class Codes: G06T 2207/10072 Tomographic images
• CPC Class Codes: G06T 2207/10116 X-ray image
• CPC Class Codes: G06T 2207/30012 Spine; Backbone
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/64 of convexity or concavity

Abstract

A method for providing automatic detection of curvature of a spine and computation of specific angles in images of the spine includes automatically displaying the curvature of the spine as a line in an image of the spine, and computing at least one of a first angle or a second angle based on the line of the curvature of the spine.

Claims

1. A method for providing automatic detection of curvature of a spine and computation of specific angles in images of the spine, comprising:

automatically displaying the curvature of the spine as a line in an image of the spine; and

computing at least one of a first angle or a second angle based on the line of the curvature of the spine.

2. The method of claim 1, wherein the first angle is a Cobb angle and the second angle is a kyphotic angle.

3. The method of claim 1, further comprising automatically detecting whether the image is a first view or a second view of the spine.

4. The method of claim 3, wherein the first view is lateral view of the spine and the second view is a coronal image of the spine.

5. The method of claim 1, wherein automatically displaying the curvature of the spine as a line in the image of the-spine comprises:

isolating the spine in the image;

improving the contrast of the image of the spine;

detecting boundaries of the spine from the improved image; and

displaying a middle line in the spine that represents the curvature of the spine in the image.

6. The method of claim 5, wherein the image is a first view of the spine, and wherein isolating the spine in the first view of the spine comprises:

determining a top cut point, the top cut point corresponding to the beginning of the spine; and

determining a bottom cut point, the bottom cut point corresponding to the bottom of the spine.

7. The method of claim 6, wherein determining the top cut point comprises detecting a head in the first view of the image.

8. The method of claim 7, wherein detecting the head in the first view of the image comprises equalizing a histogram, thresholding and segmenting the head.

9. The method of claim 6, wherein determining the bottom cut point comprises detecting a pelvis in the first view of the image.

10. The method of claim 6, wherein the first view of the image is one of a lateral view of the spine or a coronal view of the spine.

11. The method of claim 6, further comprising removing a portion of each side of the image.

12. The method of claim 5, wherein improving the contrast of the image of the spine comprises generating a mask of the spine and improving the contrast of vertebrae of the spine.

13. The method of claim 12, wherein generating a mask of the spine comprises finding a region of interest and generating a mask to enable removal of artifacts.

14. The method of claim 13, wherein finding a region of interest comprises steps of:

dividing the image into a plurality of slices or sections, wherein a width of each slice or section is equal to a width of the image, and wherein a height of each slice or section is a predetermined number of times smaller than a height of the image;

for each slice or section, equalizing a histogram to spread the intensity distribution and applying a threshold of a predetermined percentage of a maximum intensity to raise a contrast;

shifting each of the slices by one tenth of the height of the slice; and

repeating the steps of dividing, equalizing and shifting a predetermined number of times, obtaining a set of results; and

calculating a mean of the set of results.

15. The method of claim 13, wherein, when the image is a coronal view of the spine, generating the mask comprises:

generating a binary representation of the image;

applying a morphological closing to fill gaps with a small round structuring element in the binary representation of the image; and

doing a region labeling and selecting a region of interest in the binary representation of the image.

16. The method of claim 15, further comprising applying constraints in the width of the mask to improve the mask.

17. The method of claim 13, wherein, when the image is a lateral view of the spine, generating the mask comprises:

generating a binary representation of the image;

doing a region labeling and selecting a region of interest in the binary representation of the image;

separately extracting a right boundary and a left boundary of the region as vectors;

selecting the smoothest curve; and

drawing a mask following the curve and adding a predetermined value representing a width of the spine.

18. The method of claim 5, further comprising pre-processing to improve a contrast of the image before isolating the spine in the image.

19. The method of claim 5, wherein detecting boundaries of the spine from the improved image comprises:

applying a threshold to the contrast-enhanced image;

doing a region labeling of the binary image and selecting the spine;

along each row of the spine region, storing the minimum and maximum x-coordinates in two right and left vectors; and

smoothing the two vectors with the minimum and maximum x-coordinates using a smoothing window of a predetermined height.

20. The method of claim 19, wherein the predetermined height is a tenth of the image height.

21. The method of claim 5, wherein displaying a middle line in the spine that represents the curvature of the spine in the X-ray image comprises:

obtaining a middle point for each line of the spine region by computing the middle value between the pre-computed left and right vectors; and

joining all the pre computed middle points to form the representation of the curve of the spine.

22. The method of claim 1, wherein the first angle is a Cobb angle, and wherein computing the Cobb angle based on the curvature of the spine comprises:

determining the global curve of the spine;

finding the maximum positive and negative angles from the curve by calculating an angle at each point of the curve; and

computing the Cobb angle by summing the absolute value of the most negative and positive angles of the curve.

23. The method of claim 22, wherein calculating an angle at each point of the curve comprises:

computing the angle using a tangent of the curve at the specified point and a horizontal line;

locally applying a Gabor filtering along the direction set by the computed angle;

estimating an orientation from endplates of the vertebrae based on the Gabor filtering; and

setting the orientation as the Cobb angle.

24. The method of claim 1, wherein the second angle is a kyphotic angle, and wherein computing the kyphotic angle based on the curvature of the spine comprises:

determining a global curve of the spine;

finding the maximum positive and negative angles from the curve by calculating an angle at each point of the curve; and

computing the kyphotic angle by summing the absolute value of the most negative and positive angles of the curve.

25. The method of claim 24, wherein calculating an angle at each point of the curve comprises:

computing the angle using a tangent of the curve at the specified point and a horizontal line;

locally applying a Gabor filtering along the direction set by the computed angle;

estimating an orientation from endplates of the vertebrae based on the Gabor filtering; and

setting the orientation as the kyphotic angle

26. A system for providing automatic detection of curvature of a spine and computation of specific angles in images of the spine, comprising:

a memory device for storing a program;

a processor in communication with the memory device, the processor operative with the program to;

automatically display the curvature of the spine as a line in an image of the spine; and

compute at least one of a first angle or a second angle based on the line of the curvature of the spine.

27. The system of claim 26, wherein the processor is further operative with the program code to:

isolate the spine in the image;

improve the contrast of the image of the spine;

detect boundaries of the spine from the improved image; and

display a middle line in the spine that represents the curvature of the spine in the image.

28. The system of claim 26, wherein the processor is further operative with the program code to automatically detect whether the image is a lateral view of the spine or a coronal image of the spine.

29. The system of claim 26, wherein the processor is further operative with the program code to:

divide the image into a plurality of slices or sections, wherein a width of each slice or section is equal to a width of the image, and wherein a height of each slice or section is a predetermined number of times smaller than a height of the image;

for each slice or section, equalize a histogram to spread the intensity distribution and apply a threshold of a predetermined percentage of a maximum intensity to raise a contrast;

shift each of the slices by one tenth of the height of the slice; and

repeat the steps of dividing, equalizing and shifting a predetermined number of times to obtain a set of results; and

calculate a mean of the set of results.

30. The system of claim 26, wherein the processor is further operative with the program code to:

generate a binary representation of the image;

do a region labeling and select a region of interest in the binary representation of the image;

separately extract a right boundary and a left boundary of the region as vectors;

select the smoothest curve; and

draw a mask following the curve and add a predetermined value representing a width of the spine.

31. A method for providing automatic detection of curvature of a spine and computation of one of a Cobb angle or a kyphotic angle in an X-ray image of the spine, comprising:

automatically detecting whether the X-ray image is a lateral view of the spine or a coronal view of the spine;

detecting the spine in the X-ray image of the spine;

determining the curvature of the spine in the X-ray image of the spine; and

when it is determined that the X-ray image is a coronal view of the spine, computing the Cobb angle, and when it is determined that the X-ray image is a lateral view of the spine, computing the kyphotic angle.