Method and system for locating opaque regions in chest X-ray radiographs

1. A method for segmenting a heart region in a chest X-ray radiograph, comprising:

generating a heart region boundary based on lung boundaries in the chest X-ray radiograph and an average heart region model;

determining a location of a lower boundary of the heart region boundary;

registering left and right portions of the heart region boundary to corresponding portions of the lung boundaries; and

adjusting upper and lower portions of the heart region boundary based on the left and right portions of the heart region boundary.

2. The method of claim 1 , wherein said step of generating a heart region boundary comprises:

detecting a corner point on each of left and right lung boundaries;

estimating the heart region boundary by aligning points on the average heart model with the corner points on the left and right lung boundaries;

scaling the heart region boundary based on a distance between the corner points on the left and right lung boundaries; and

tuning a vertical position of the heart region boundary based on overlap between the heart region boundary and the left and right lung boundaries.

3. The method of claim 1 , wherein said step of determining a location of a lower boundary of the heart region boundary comprises:

determining the location of said lower boundary by detecting lines in said chest X-ray radiograph using a Hough transform.

4. The method of claim 1 wherein said step of determining a location of a lower boundary of the heart region boundary comprises:

generating a gradient image in a horizontal direction of a region of interest of said chest X-ray radiograph, said region of interest including a lower portion of the heart region boundary;

generating a binary image of said gradient image by binarizing said gradient image;

thinning said binary image;

detecting one or more candidate lines using a Hough transform of the thinned binary image; and

selecting one of the candidate lines as the lower boundary of the heart region boundary.

5. The method of claim 4 , wherein said step of detecting one or more candidate lines using a Hough transform of the thinned binary image comprises:

detecting lines in the thinned binary image using the Hough transform; and

determining a line to be a candidate line when the line is longer than a threshold.

6. The method of claim 4 , wherein said step of selecting one of the candidate lines as the lower boundary of the heart region comprises:

selecting a lowest candidate line in the thinned binary image as the lower boundary of the heart region.

7. The method of claim 6 , wherein said step of selecting one of the candidate lines as the lower boundary of the heart region further comprises:

generating a new binary image of said gradient image by binarizing said gradient image using a higher binarization threshold than said binary image;

thinning said new binary image;

determining whether any candidate lines are detected in the thinned new binary image using a Hough transform;

if one or more candidate lines are detected in the thinned new binary image, selecting a lowest candidate line in the thinned binary image as the lower boundary of the heart region.

8. The method of claim 4 , wherein said step of determining a location of a lower boundary of the heart region boundary further comprises:

adjusting each point in the lower boundary of the heart region boundary to minimize an energy function based on an intensity difference between an inside and an outside of the heart region boundary, a gradient at each point, and an intensity difference between the inside of the heart region boundary and a center of the heart region boundary.

9. The method of claim 1 , wherein said step of registering left and right portions of the heart region boundary to corresponding portions of the lung boundaries comprises:

determining intersection points on each of left and right lung boundaries that correspond to points on the left and right portions of the heart region boundary, respectively; and

calculating registration parameters to transform the points on each of the left and right portions of the heart boundary to the corresponding points on each of the left and right lung boundaries, respectively.

10. The method of claim 1 , wherein said step of adjusting upper and lower portions of the heart region boundary based on the left and right portions of the heart region boundary comprises:

at least one of shrinking, enlarging, and rotating said upper and lower portions of the heart region boundary such that the upper, lower, left and right portions of the heart region boundary form a smooth contour.

11. The method of claim 1 , further comprising:

outputting the heart region boundary.

12. The method of claim 1 , further comprising:

detecting nodules within the heart region boundary in said chest X-ray radiograph separately from lung regions in said chest X-ray radiograph.

13. An apparatus for segmenting a heart region in a chest X-ray radiograph, comprising:

means for generating a heart region boundary based on lung boundaries in the chest X-ray radiograph and an average heart region model;

means for determining a location of a lower boundary of the heart region boundary;

means for registering left and right portions of the heart region boundary to corresponding portions of the lung boundaries; and

means for adjusting upper and lower portions of the heart region boundary based on the left and right portions of the heart region boundary.

14. The apparatus of claim 13 , wherein said means for generating a heart region boundary comprises:

means for detecting a corner point on each of left and right lung boundaries;

means for estimating the heart region boundary by aligning points on the average heart model with the corner points on the left and right lung boundaries;

means for scaling the heart region boundary based on a distance between the corner points on the left and right lung boundaries; and

means for tuning a vertical position of the heart region boundary based on overlap between the heart region boundary and the left and right lung boundaries.

15. The apparatus of claim 13 , wherein said means for determining a location of a lower boundary of the heart region boundary comprises:

means for generating a gradient image in a horizontal direction of a region of interest of said chest X-ray radiograph, said region of interest including a lower portion of the heart region boundary;

means for generating a binary image of said gradient image by binarizing said gradient image;

means for thinning said binary image;

means for detecting one or more candidate lines using a Hough transform of the thinned binary image; and

means for selecting one of the candidate lines as the lower boundary of the heart region boundary.

16. The apparatus of claim 15 , wherein said means for selecting one of the candidate lines as the lower boundary of the heart region comprises:

means selecting a lowest candidate line in the thinned binary image as the lower boundary of the heart region.

17. The apparatus of claim 15 , wherein said means for determining a location of a lower boundary of the heart region boundary further comprises:

means for adjusting each point in the lower boundary of the heart region boundary to minimize an energy function based on an intensity difference between an inside and an outside of the heart region boundary, a gradient at each point, and an intensity difference between the inside of the heart region boundary and a center of the heart region boundary.

18. The apparatus of claim 13 , wherein said means for registering left and right portions of the heart region boundary to corresponding portions of the lung boundaries comprises:

means for determining intersection points on each of left and right lung boundaries that correspond to points on the left and right portions of the heart region boundary, respectively; and

means for calculating registration parameters to transform the points on each of the left and right portions of the heart boundary to the corresponding points on each of the left and right lung boundaries, respectively.

19. A computer readable medium encoded with computer executable instructions for segmenting a heart region in a chest X-ray radiograph, the computer executable instructions defining steps comprising:

generating a heart region boundary based on lung boundaries in the chest X-ray radiograph and an average heart region model;

determining a location of a lower boundary of the heart region boundary;

registering left and right portions of the heart region boundary to corresponding portions of the lung boundaries; and

adjusting upper and lower portions of the heart region boundary based on the left and right portions of the heart region boundary.

20. The computer readable medium of claim 19 , wherein said the computer executable instructions defining the step of generating a heart region boundary comprise computer executable instructions defining the steps of:

detecting a corner point on each of left and right lung boundaries;

estimating the heart region boundary by aligning points on the average heart model with the corner points on the left and right lung boundaries;

scaling the heart region boundary based on a distance between the corner points on the left and right lung boundaries; and

tuning a vertical position of the heart region boundary based on overlap between the heart region boundary and the left and right lung boundaries.

21. The computer readable medium of claim 19 , wherein the computer executable instructions defining the step of determining a location of a lower boundary of the heart region boundary comprise computer executable instructions defining the steps of:

generating a gradient image in a horizontal direction of a region of interest of said chest X-ray radiograph, said region of interest including a lower portion of the heart region boundary;

generating a binary image of said gradient image by binarizing said gradient image;

thinning said binary image;

detecting one or more candidate lines using a Hough transform of the thinned binary image; and

selecting one of the candidate lines as the lower boundary of the heart region boundary.

22. The computer readable medium of claim 21 , wherein the computer executable instructions defining the step of detecting one or more candidate lines using a Hough transform of the thinned binary image comprise computer executable instructions defining the steps of:

detecting lines in the thinned binary image using the Hough transform; and

determining a line to be a candidate line when the line is longer than a threshold.

23. The computer readable medium of claim 21 , wherein the computer executable instructions defining the step of selecting one of the candidate lines as the lower boundary of the heart region comprise computer executable instructions defining the step of:

selecting a lowest candidate line in the thinned binary image as the lower boundary of the heart region.

24. The computer readable medium of claim 21 , wherein the computer executable instructions defining the step of determining a location of a lower boundary of the heart region boundary further comprise computer executable instructions defining the step of:

adjusting each point in the lower boundary of the heart region boundary to minimize an energy function based on an intensity difference between an inside and an outside of the heart region boundary, a gradient at each point, and an intensity difference between the inside of the heart region boundary and a center of the heart region boundary.

25. The computer readable medium of claim 19 , wherein the computer executable instructions defining the step of registering left and right portions of the heart region boundary to corresponding portions of the lung boundaries comprise computer executable instructions defining the steps of:

determining intersection points on each of left and right lung boundaries that correspond to points on the left and right portions of the heart region boundary, respectively; and

calculating registration parameters to transform the points on each of the left and right portions of the heart boundary to the corresponding points on each of the left and right lung boundaries, respectively.