Method and system for diaphragm segmentation in chest X-ray radiographs

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

segmenting a right diaphragm region based on a boundary of a right lung lobe region in the chest X-ray radiograph; and

segmenting a left diaphragm region based on a heart region and a boundary of a left lung lobe region in the chest X-ray radiograph.

2. The method of claim 1 , wherein said step of segmenting a right diaphragm region comprises:

detecting an upper boundary of the right diaphragm region coincident with a lower boundary of the right lung lobe region; and

determining a lower boundary of the right diaphragm region as a symmetric reflection of the upper boundary of the right diaphragm region.

3. The method of claim 2 , wherein said step of segmenting a right diaphragm region further comprises:

adjusting points on the lower boundary of the right diaphragm region based on local edge information.

4. The method of claim 1 , wherein said step of segmenting a left diaphragm region comprises:

detecting an inner corner of the left diaphragm region;

detecting an outer corner of the left diaphragm region;

detecting an inner upper boundary of the left diaphragm region extending from the inner corner of the left diaphragm region;

detecting an outer upper boundary of the left diaphragm region extending from the outer corner of the left diaphragm region and connecting with the inner upper boundary of the left diaphragm region to form an upper boundary of the left diaphragm region; and

determining a lower boundary of the left diaphragm region based on the inner and outer corners of the left diaphragm region.

5. The method of claim 4 , wherein said step of detecting an inner corner of the left diaphragm region comprises:

applying vertical edge enhancement inside a region of interest defined by the heart region resulting in an edge enhanced image;

binarizing and thinning the edge enhance image;

detecting connect components within the region of interest;

determining whether a largest one the connected components is longer than a threshold value;

if the largest one of the connected components is not longer than the threshold value, using a default point on a boundary of the heart region as the inner corner of the left diaphragm; and

if the largest one of the connected components is longer than the threshold value,

selecting an optimal connect component,

determining an inner vertical boundary by fitting a curve to the selected optimal connected component, and

detecting the inner corner as an intersection of the inner vertical boundary and a lower boundary of the heart region.

6. The method of claim 4 , wherein said step of detecting an outer corner of the left diaphragm region further comprises:

detecting an outer vertical boundary in a region of interest defined as a lower portion of the left lung lobe region below an upper boundary of the heart region; and

detecting an intersection between the outer vertical boundary and a lower boundary of the left lung lobe region.

7. The method of claim 4 , wherein said step of detecting an inner upper boundary of the left diaphragm region comprises:

applying horizontal edge enhancement inside a region of interest defined by the heart region resulting in an edge enhanced image;

binarizing and thinning the edge enhanced image at a first binarization threshold;

detecting connected components larger than a threshold value in the region of interest;

selecting an optimal connected component;

removing connected components other than the optimal connected component and re-binarizing and thinning the edge enhanced image at a second binarization threshold;

re-calculating the optimal connected component;

detecting the inner upper boundary of the left diaphragm region by fitting a curve to the optimal connected component.

8. The method of claim 4 , wherein said step of detecting an outer upper boundary of the left diaphragm region comprises:

detecting an outer upper boundary of the left diaphragm region coincident with a lower boundary of the left lung lobe.

9. The method of claim 4 , wherein said step of determining a lower boundary of the left diaphragm region based on the inner and outer corners of the left diaphragm region comprises:

estimating an ellipsoid shape that passes through the inner and outer corners of the left diaphragm region.

10. The method of claim 4 , wherein said step of segmenting a left diaphragm region further comprises:

adjusting points on the lower boundary of the left diaphragm region based on local edge information.

11. An apparatus for segmenting diaphragm regions in a chest X-ray radiograph, comprising:

means for segmenting a right diaphragm region based on a boundary of a right lung lobe region in the chest X-ray radiograph; and

means for segmenting a left diaphragm region based on a heart region and a boundary of a left lung lobe region in the chest X-ray radiograph.

12. The apparatus of claim 11 , wherein said means for segmenting a right diaphragm region comprises:

means for detecting an upper boundary of the right diaphragm region coincident with a lower boundary of the right lung lobe region; and

means for determining a lower boundary of the right diaphragm region as a symmetric reflection of the upper boundary of the right diaphragm region.

13. The apparatus of claim 1 , wherein said means for segmenting a left diaphragm region comprises:

means for detecting an inner corner of the left diaphragm region;

means for detecting an outer corner of the left diaphragm region;

means for detecting an inner upper boundary of the left diaphragm region extending from the inner corner of the left diaphragm region;

means for detecting an outer upper boundary of the left diaphragm region extending from the outer corner of the left diaphragm region and connecting with the inner upper boundary of the left diaphragm region to form an upper boundary of the left diaphragm region; and

means for determining a lower boundary of the left diaphragm region based on the inner and outer corners of the left diaphragm region.

14. The apparatus of claim 13 , wherein said means for detecting an inner corner of the left diaphragm region comprises:

means for applying vertical edge enhancement inside a region of interest defined by the heart region resulting in an edge enhanced image;

means for binarizing and thinning the edge enhance image;

means for detecting connect components within the region of interest;

means for selecting an optimal connect component;

means for determining an inner vertical boundary by fitting a curve to the selected optimal connected component; and

means for detecting the inner corner as an intersection of the inner vertical boundary and a lower boundary of the heart region.

15. The apparatus of claim 13 , wherein said means for detecting an outer corner of the left diaphragm region further comprises:

means for detecting an outer vertical boundary in a region of interest defined as a lower portion of the left lung lobe region below an upper boundary of the heart region; and

means for detecting an intersection between the outer vertical boundary and a lower boundary of the left lung lobe region.

16. The apparatus of claim 13 , wherein said means for detecting an inner upper boundary of the left diaphragm region comprises:

means for applying horizontal edge enhancement inside a region of interest defined by the heart region resulting in an edge enhanced image;

means for binarizing and thinning the edge enhanced image;

means for detecting connected components larger than a threshold value in the region of interest;

means for selecting an optimal connected component; and

means for detecting the inner upper boundary of the left diaphragm region by fitting a curve to the optimal connected component.

17. The apparatus of claim 13 , wherein said means for detecting an outer upper boundary of the left diaphragm region comprises:

means for detecting an outer upper boundary of the left diaphragm region coincident with a lower boundary of the left lung lobe.

18. The method of claim 4 , wherein said means for determining a lower boundary of the left diaphragm region based on the inner and outer corners of the left diaphragm region comprises:

means for estimating an ellipsoid shape that passes through the inner and outer corners of the left diaphragm region.

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

segmenting a right diaphragm region based on a boundary of a right lung lobe region in the chest X-ray radiograph; and

segmenting a left diaphragm region based on a heart region and a boundary of a left lung lobe region in the chest X-ray radiograph.

20. The computer readable medium of claim 19 , wherein the computer executable instructions defining the step of segmenting a right diaphragm region comprise computer executable instructions defining the steps of:

detecting an upper boundary of the right diaphragm region coincident with a lower boundary of the right lung lobe region; and

determining a lower boundary of the right diaphragm region as a symmetric reflection of the upper boundary of the right diaphragm region.

21. The computer readable medium of claim 19 , wherein the computer executable instructions defining the step of segmenting a left diaphragm region comprise computer executable instructions defining the steps of:

detecting an inner corner of the left diaphragm region;

detecting an outer corner of the left diaphragm region;

detecting an inner upper boundary of the left diaphragm region extending from the inner corner of the left diaphragm region;

detecting an outer upper boundary of the left diaphragm region extending from the outer corner of the left diaphragm region and connecting with the inner upper boundary of the left diaphragm region to form an upper boundary of the left diaphragm region; and

determining a lower boundary of the left diaphragm region based on the inner and outer corners of the left diaphragm region.

22. The computer readable medium of claim 21 , wherein the computer executable instructions defining the step of detecting an inner corner of the left diaphragm region comprise computer executable instructions defining the steps of:

applying vertical edge enhancement inside a region of interest defined by the heart region resulting in an edge enhanced image;

binarizing and thinning the edge enhance image;

detecting connect components within the region of interest;

determining whether a largest one the connected components is longer than a threshold value;

if the largest one of the connected components is not longer than the threshold value, using a default point on a boundary of the heart region as the inner corner of the left diaphragm; and

if the largest one of the connected components is longer than the threshold value,

selecting an optimal connect component,

determining an inner vertical boundary by fitting a curve to the selected optimal connected component, and

detecting the inner corner as an intersection of the inner vertical boundary and a lower boundary of the heart region.

23. The computer readable medium of claim 21 , wherein the computer executable instructions defining the step of detecting an outer corner of the left diaphragm region further comprise computer executable instructions defining the steps of:

detecting an outer vertical boundary in a region of interest defined as a lower portion of the left lung lobe region below an upper boundary of the heart region; and

detecting an intersection between the outer vertical boundary and a lower boundary of the left lung lobe region.

24. The computer readable medium of claim 21 , wherein the computer executable instructions defining the step of detecting an inner upper boundary of the left diaphragm region comprise computer executable instructions defining the steps of:

applying horizontal edge enhancement inside a region of interest defined by the heart region resulting in an edge enhanced image;

binarizing and thinning the edge enhanced image at a first binarization threshold;

detecting connected components larger than a threshold value in the region of interest;

selecting an optimal connected component;

removing connected components other than the optimal connected component and re-binarizing and thinning the edge enhanced image at a second binarization threshold;

re-calculating the optimal connected component; and

detecting the inner upper boundary of the left diaphragm region by fitting a curve to the optimal connected component.

25. The computer readable medium of claim 21 , wherein the computer executable instructions defining the step of detecting an outer upper boundary of the left diaphragm region comprise computer executable instructions defining the step of:

detecting an outer upper boundary of the left diaphragm region coincident with a lower boundary of the left lung lobe.

26. The computer readable medium of claim 21 , wherein the computer executable instructions defining the step of determining a lower boundary of the left diaphragm region based on the inner and outer corners of the left diaphragm region comprise computer executable instructions defining the step of:

estimating an ellipsoid shape that passes through the inner and outer corners of the left diaphragm region.