System and method for coronary digital subtraction angiography

1. A method for extracting coronary vessels from a contrast image, comprising:

receiving a plurality of mask images of a coronary region;

receiving a contrast image of the coronary region;

detecting vessel regions in the contrast image using learning-based vessel segment detection;

determining a background region of the contrast image based on the detected vessel regions;

estimating background motion between one of the plurality of mask images and the background region of the contrast image by estimating a motion field between the one of the plurality of mask images and the background image and performing covariance-based filtering over the estimated motion field;

warping said one of the plurality of mask images based on the estimated background motion to generate an estimated background layer; and

extracting a coronary vessel layer from the contrast image by subtracting the estimated background layer from the contrast image.

2. The method of claim 1 , further comprising:

repeating said steps of receiving a contrast image, detecting vessel regions in the contrast image, determining a background region of the contrast image, estimating background motion, warping one of the plurality of mask images, and extracting a coronary vessel layer from the contrast image for each of a sequence of contrast images.

3. The method of claim 2 , wherein the coronary vessel layer for each of the sequence of contrast images is extracted in real-time as each of the sequence of contrast images is received.

4. The method of claim 1 , wherein the plurality of mask images comprises a sequence of mask images of the coronary region taken over at least one full cardiac cycle.

5. The method of claim 1 , wherein said step of detecting vessel regions in the contrast image comprises:

applying a set of steerable filters to the contrast image to detect ridge features in the contrast image; and

detecting vessel segments at the detected ridge features using a learned binary classifier.

6. The method of claim 1 , wherein said step of estimating background motion between one of the plurality of mask images and the background region of the contrast image comprises:

estimating the motion field between each mask image and the background region of the contrast image; and

selecting one of the plurality of mask images that best matches the background region of the contrast image based on the estimated motion fields between each of the plurality of mask images and the background region of the contrast image.

7. The method of claim 1 , wherein said step of estimating background motion between one of the plurality of mask images and the background region of the contrast image comprises:

(a) estimating an initial motion field between the one of the plurality of mask images and the background region of the contrast image and covariance matrices of estimated motion vectors in the motion field;

(b) warping the one of the plurality of mask images based on the estimated motion field;

(c) calculating an error image between the background region of the contrast image and the warped one of the plurality of mask images;

(d) calculating an image gradient of the warped one of the plurality of mask images;

(e) calculating increment motion vectors based on the image gradient and covariance matrices of the increment motion vectors;

(f) calculated adjusted increment motion vectors by filtering the motion field based on the covariance matrices; and

(g) updating the estimated motion field based on the adjusted increment motion vectors.

8. The method of claim 7 , wherein said step of estimating background motion between one of the plurality of mask images and the background region of the contrast image further comprises:

(h) repeating steps (b)-(g) until convergence.

9. The method of claim 8 , wherein said step of estimating background motion between one of the plurality of mask images and the background region of the contrast image further comprises:

performing steps (a)-(h) for each of a plurality of scales from coarsest to finest.

10. The method of claim 1 , further comprising:

generating an enhanced coronary image by enhancing coronary vessels in the contrast image based on the extracted coronary vessel layer.

11. An apparatus for extracting coronary vessels from a contrast image, comprising:

means for receiving a plurality of mask images of a coronary region;

means for receiving a contrast image of the coronary region;

means for detecting vessel regions in the contrast image using learning-based vessel segment detection;

means for determining a background region of the contrast image based on the detected vessel regions;

means for estimating background motion between one of the plurality of mask images and the background region of the contrast image by estimating a motion field between the one of the plurality of mask images and the background image and performing covariance-based filtering over the estimated motion field;

means for warping said one of the plurality of mask images based on the estimated background motion to generate an estimated background layer; and

means for extracting a coronary vessel layer from the contrast image by subtracting the estimated background layer from the contrast image.

12. The apparatus of claim 11 , wherein said means for receiving a contrast image of the coronary region comprises:

means for receiving a sequence of contrast images of the coronary region.

13. The apparatus of claim 11 , wherein the plurality of mask images comprises a sequence of mask images of the coronary region taken over at least one full cardiac cycle.

14. The apparatus of claim 11 , wherein said means for detecting vessel regions in the contrast image comprises:

means for applying a set of steerable filters to the contrast image to detect ridge features in the contrast image; and

means for detecting vessel segments at the detected ridge features using a learned binary classifier.

15. The apparatus of claim 11 , wherein said means for estimating background motion between one of the plurality of mask images and the background region of the contrast image comprises:

means for estimating the motion field between each mask image and the background region of the contrast image; and

means for selecting one of the plurality of mask images that best matches the background region of the contrast image based on the estimated motion fields between each of the plurality of mask images and the background region of the contrast image.

16. The apparatus of claim 11 , wherein said means for estimating background motion between one of the plurality of mask images and the background region of the contrast image comprises:

means for estimating an initial motion field between the one of the plurality of mask images and the background region of the contrast image;

means for warping the one of the plurality of mask images based on the estimated motion field;

means for calculating an error image between the background region of the contrast image and the warped one of the plurality of mask images;

means for calculating an image gradient of the warped one of the plurality of mask images;

means for calculating increment motion vectors based on the image gradient and covariance matrices of the increment motion vectors;

means for calculated adjusted increment motion vectors by filtering the motion field based on the covariance matrices; and

means for updating the estimated motion field based on the adjusted increment motion vectors.

17. The apparatus of claim 9 , further comprising:

means for generating an enhanced coronary image by enhancing coronary vessels in the contrast image based on the extracted coronary vessel layer.

18. A non-transitory computer readable medium encoded with computer executable instructions for extracting for extracting coronary vessels from a contrast image, the computer executable instructions defining steps comprising:

receiving a plurality of mask images of a coronary region;

receiving a contrast image of the coronary region;

detecting vessel regions in the contrast image using learning-based vessel segment detection;

determining a background region of the contrast image based on the detected vessel regions;

estimating background motion between one of the plurality of mask images and the background region of the contrast image by estimating a motion field between the one of the plurality of mask images and the background image and performing covariance-based filtering over the estimated motion field;

warping said one of the plurality of mask images based on the estimated background motion to generate an estimated background layer; and

extracting a coronary vessel layer from the contrast image by subtracting the estimated background layer from the contrast image.

19. The non-transitory computer readable medium of claim 18 , further comprising computer executable instructions defining the step of:

repeating said steps of receiving a contrast image, detecting vessel regions in the contrast image, determining a background region of the contrast image, estimating background motion, warping one of the plurality of mask images, and extracting a coronary vessel layer from the contrast image for each of a sequence of contrast images.

20. The non-transitory computer readable medium of claim 19 , wherein the coronary vessel layer for each of the sequence of contrast images is extracted in real-time as each of sequence of contrast images is received.

21. The non-transitory computer readable medium of claim 18 , wherein the plurality of mask images comprises a sequence of mask images of the coronary region taken over at least one full cardiac cycle.

22. The non-transitory computer readable medium of claim 18 , wherein the computer executable instructions defining the step of detecting vessel regions in the contrast image comprise computer executable instructions defining the steps of:

applying a set of steerable filters to the contrast image to detect ridge features in the contrast image; and

detecting vessel segments at the detected ridge features using a learned binary classifier.

23. The non-transitory computer readable medium of claim 18 , wherein the computer executable instructions defining the step of estimating background motion between one of the plurality of mask images and the background region of the contrast image comprise computer executable instructions defining the steps of:

estimating the motion field between each mask image and the background region of the contrast image; and

selecting one of the plurality of mask images that best matches the background region of the contrast image based on the estimated motion fields between each of the plurality of mask images and the background region of the contrast image.

24. The non-transitory computer readable medium of claim 18 , wherein the computer executable instructions defining the step of estimating background motion between one of the plurality of mask images and the background region of the contrast image comprise computer executable instructions defining the steps of:

(a) estimating an initial motion field between the one of the plurality of mask images and the background region of the contrast image and covariance matrices of estimated motion vectors in the motion field;

(b) warping the one of the plurality of mask images based on the estimated motion field;

(c) calculating an error image between the background region of the contrast image and the warped one of the plurality of mask images;

(d) calculating an image gradient of the warped one of the plurality of mask images;

(e) calculating increment motion vectors based on the image gradient and covariance matrices of the increment motion vectors;

(f) calculated adjusted increment motion vectors by filtering the motion field based on the covariance matrices; and

(g) updating the estimated motion field based on the adjusted increment motion vectors.

25. The non-transitory computer readable medium of claim 24 , wherein the computer executable instructions defining the step of estimating background motion between one of the plurality of mask images and the background region of the contrast image further comprise computer executable instructions defining the step of:

(h) repeating steps (b)-(g) until convergence.

26. The non-transitory computer readable medium of claim 18 , wherein the computer executable instructions defining the step of estimating background motion between one of the plurality of mask images and the background region of the contrast image further comprise computer executable instructions defining the step of:

performing steps (a)-(h) for each of a plurality of scales from coarsest to finest.

27. The non-transitory computer readable medium of claim 18 , further comprising computer executable instructions defining the step of:

generating an enhanced coronary image by enhancing coronary vessels in the contrast image based on the extracted coronary vessel layer.