Method and system for guiding catheter detection in fluoroscopic images

1. A method for detecting a guiding catheter in a fluoroscopic image, comprising:

detecting a plurality of guiding catheter centerline segment candidates in the fluoroscopic image;

clustering the plurality of guiding catheter centerline segment candidates into a number of guiding catheter centerline segment clusters; and

detecting a guiding catheter centerline connecting a guiding catheter centerline ending point in the fluoroscopic image with an image margin of the fluoroscopic image by connecting a plurality of the guiding catheter centerline segment clusters.

2. The method of claim 1 , further comprising:

automatically detecting the image margin of the fluoroscopic image; and

generating a ridge map of the fluoroscopic image by ridge filtering the fluoroscopic image, where the step of detecting a plurality of guiding catheter centerline segments is constrained based on the detected image margin and the ridge map.

3. The method of claim 2 , wherein the step of automatically detecting the image margin of the fluoroscopic image comprises:

detecting the image margin of the fluoroscopic image using mean thresholding.

4. The method of claim 2 , wherein the step of generating a ridge map of the fluoroscopic image by ridge filtering the fluoroscopic image comprises:

down-sampling the fluoroscopic image; and

filtering the down-sampled fluoroscopic image using ridge filtering.

5. The method of claim 1 , wherein the step of detecting a plurality of guiding catheter centerline segment candidates in the fluoroscopic image comprises:

detecting the plurality of guiding catheter centerline segment candidates in the fluoroscopic image using a guiding catheter centerline segment detector trained based on annotated training data.

6. The method of claim 5 , wherein the guiding catheter centerline segment detector is a probabilistic boosting tree (PBT) trained using 2D Haar-like features.

7. The method of claim 1 , wherein the step of detecting a guiding catheter centerline connecting a guiding catheter centerline ending point in the fluoroscopic image with an image margin of the fluoroscopic image based on the plurality of guiding catheter centerline segment candidates comprises:

detecting the guiding catheter centerline connecting the guiding catheter centerline ending point with the image margin in a guiding catheter centerline ending direction.

8. The method of claim 1 , wherein the step of detecting a guiding catheter centerline connecting a guiding catheter centerline ending point in the fluoroscopic image with an image margin of the fluoroscopic image based on the plurality of guiding catheter centerline segment candidates comprises:

generating a graph of the guiding catheter centerline segment candidates;

generating a tree of shortest path candidates based on the graph; and

selecting one of the shortest path candidates as the guiding catheter centerline.

9. The method of claim 8 , wherein the step of generating a graph of the guiding catheter centerline segment candidates comprises:

representing each of the plurality of guiding catheter centerline segment candidates as a node in the graph;

generating edges connecting the node representing each guiding catheter centerline segment candidate with one or more nodes representing at most a predetermined number of spatially closest guiding catheter centerline segment candidates; and

calculating a pair-wise connectivity score for each edge as a sum of absolute pixel intensity differences in the fluoroscopic image along a spline arc connecting the guiding catheter centerline segment candidates represented by the nodes connected by the edge.

10. The method of claim 8 , wherein the step of generating a tree of shortest path candidates based on the graph comprises:

generating a shortest path from a node in the graph representing the guiding catheter centerline ending point through each of a plurality of nodes in the graph representing guiding catheter centerline segment candidates spatially closest to the guiding catheter centerline ending point.

11. The method of claim 8 , wherein the step of selecting one of the shortest path candidates as the guiding catheter centerline comprises:

selecting a shortest one of the shortest path candidates that connects the guiding catheter centerline ending point with the image margin and that has an orientation corresponding to a guiding catheter ending direction.

12. An apparatus for detecting a guiding catheter in a fluoroscopic image, comprising:

means for detecting a plurality of guiding catheter centerline segment candidates in the fluoroscopic image;

means for clustering the plurality of guiding catheter centerline segment candidates into a number of guiding catheter centerline segment clusters; and

means for detecting a guiding catheter centerline connecting a guiding catheter centerline ending point in the fluoroscopic image with an image margin of the fluoroscopic image by connecting a plurality of the guiding catheter centerline segment clusters.

13. The apparatus of claim 12 , further comprising:

means for automatically detecting the image margin of the fluoroscopic image; and

means for generating a ridge map of the fluoroscopic image by ridge filtering the fluoroscopic image, where the detecting a plurality of guiding catheter centerline segments is constrained based on the detected image margin and the ridge map.

14. The apparatus of claim 12 , wherein the means for detecting a plurality of guiding catheter centerline segment candidates in the fluoroscopic image comprises:

means for detecting the plurality of guiding catheter centerline segment candidates in the fluoroscopic image using a guiding catheter centerline segment detector trained based on annotated training data.

15. The apparatus of claim 12 , wherein the means for detecting a guiding catheter centerline connecting a guiding catheter centerline ending point in the fluoroscopic image with an image margin of the fluoroscopic image based on the plurality of guiding catheter centerline segment candidates comprises:

means for detecting the guiding catheter centerline connecting the guiding catheter centerline ending point with the image margin in a guiding catheter centerline ending direction.

16. The apparatus of claim 15 , wherein the means for detecting a guiding catheter centerline connecting a guiding catheter centerline ending point in the fluoroscopic image with an image margin of the fluoroscopic image based on the plurality of guiding catheter centerline segment candidates comprises:

means for generating a graph of the guiding catheter centerline segment candidates;

means for generating a tree of shortest path candidates based on the graph; and

means for selecting one of the shortest path candidates as the guiding catheter centerline.

17. The apparatus of claim 16 , wherein the means for generating a graph of the guiding catheter centerline segment candidates comprises:

means for representing each of the plurality of guiding catheter centerline segment candidates as a node in the graph;

means for generating edges connecting the node representing each guiding catheter centerline segment candidate with one or more nodes representing at most a predetermined number of spatially closest guiding catheter centerline segment candidates; and

means for calculating a pair-wise connectivity score for each edge as a sum of absolute pixel intensity differences in the fluoroscopic image along a spline arc connecting the guiding catheter centerline segment candidates represented by the nodes connected by the edge.

18. The apparatus of claim 16 , wherein the means for generating a tree of shortest path candidates based on the graph comprises:

means for generating a shortest path from a node in the graph representing the guiding catheter centerline ending point through each of a plurality of nodes in the graph representing guiding catheter centerline segment candidates spatially closest to the guiding catheter centerline ending point.

19. The apparatus of claim 16 , wherein the means for selecting one of the shortest path candidates as the guiding catheter centerline comprises:

means for selecting a shortest one of the shortest path candidates that connects the guiding catheter centerline ending point with the image margin and that has an orientation corresponding to a guiding catheter ending direction.

20. A non-transitory computer readable medium encoded with computer executable instructions for detecting a guiding catheter in a fluoroscopic image, the computer executable instructions defining steps comprising:

detecting a plurality of guiding catheter centerline segment candidates in the fluoroscopic image;

clustering the plurality of guiding catheter centerline segment candidates into a number of guiding catheter centerline segment clusters; and

detecting a guiding catheter centerline connecting a guiding catheter centerline ending point in the fluoroscopic image with an image margin of the fluoroscopic image by connecting a plurality of the guiding catheter centerline segment clusters.

21. The computer readable medium of claim 20 , further comprising computer executable instructions defining the steps of:

automatically detecting the image margin of the fluoroscopic image; and

generating a ridge map of the fluoroscopic image by ridge filtering the fluoroscopic image, where the step of detecting a plurality of guiding catheter centerline segments is constrained based on the detected image margin and the ridge map.

22. The computer readable medium of claim 20 , wherein the computer executable instructions defining the step of detecting a plurality of guiding catheter centerline segment candidates in the fluoroscopic image comprise computer executable instructions defining the step of:

detecting the plurality of guiding catheter centerline segment candidates in the fluoroscopic image using a guiding catheter centerline segment detector trained based on annotated training data.

23. The computer readable medium of claim 20 , wherein the computer executable instructions defining the step of detecting a guiding catheter centerline connecting a guiding catheter centerline ending point in the fluoroscopic image with an image margin of the fluoroscopic image based on the plurality of guiding catheter centerline segment candidates comprise computer executable instructions defining the step of:

detecting the guiding catheter centerline connecting the guiding catheter centerline ending point with the image margin in a guiding catheter centerline ending direction.

24. The computer readable medium of claim 20 , wherein the computer executable instructions defining the step of detecting a guiding catheter centerline connecting a guiding catheter centerline ending point in the fluoroscopic image with an image margin of the fluoroscopic image based on the plurality of guiding catheter centerline segment candidates comprise computer executable instructions defining the steps of:

generating a graph of the guiding catheter centerline segment candidates;

generating a tree of shortest path candidates based on the graph; and

selecting one of the shortest path candidates as the guiding catheter centerline.

25. The computer readable medium of claim 24 , wherein the computer executable instructions defining the step of generating a graph of the guiding catheter centerline segment candidates comprise computer executable instructions defining the steps of:

representing each of the plurality of guiding catheter centerline segment candidates as a node in the graph;

generating edges connecting the node representing each guiding catheter centerline segment candidate with one or more nodes representing at most a predetermined number of spatially closest guiding catheter centerline segment candidates; and

calculating a pair-wise connectivity score for each edge as a sum of absolute pixel intensity differences in the fluoroscopic image along a spline arc connecting the guiding catheter centerline segment candidates represented by the nodes connected by the edge.

26. The computer readable medium of claim 24 , wherein the computer executable instructions defining the step of generating a tree of shortest path candidates based on the graph comprise computer executable instructions defining the step of:

generating a shortest path from a node in the graph representing the guiding catheter centerline ending point through each of a plurality of nodes in the graph representing guiding catheter centerline segment candidates spatially closest to the guiding catheter centerline ending point.

27. The computer readable medium of claim 24 , wherein the computer executable instructions defining the step of selecting one of the shortest path candidates as the guiding catheter centerline comprise computer executable instructions defining the step of:

selecting a shortest one of the shortest path candidates that connects the guiding catheter centerline ending point with the image margin and that has an orientation corresponding to a guiding catheter ending direction.