Systems and methods for segmentation of radiopaque structures in images

1. A method of determining a model of a marker, comprising:

obtaining projection images, each of the projection images having an image of a marker that indicates a shape of the marker, the marker being a physical object implanted inside a patient;

determining binary images of the marker for respective ones of the projection images; and

constructing, using a processing unit, a three-dimensional model of the marker using the binary images, the three-dimensional model comprising a set of voxels in a three-dimensional space that collectively indicates an actual three-dimensional shape of the marker.

2. The method of claim 1 , wherein the set of voxels for the three-dimensional model of the marker comprises a first voxel and a second voxel, wherein the first voxel has a first value associated therewith to indicate a first degree of certainty that the first voxel belongs to the three-dimensional model, and the second voxel has a second value associated therewith to indicate a second degree of certainty that the second voxel belongs to the three-dimensional model, the first value and the second value being different.

3. The method of claim 1 , wherein the obtaining the projection images comprises:

receiving a set of reference images; and

selecting a subset of the reference images as the projection images;

wherein two of the reference images in the selected subset have an angular separation that is at least 15 degrees; and

wherein each of the two of the reference images in the selected subset has a marker contrast that is higher than a prescribed value.

4. A method for determining marker position, comprising creating a template using the three-dimensional model of the marker of claim 1 .

5. The method of claim 4 , wherein the template is created by projecting the three-dimensional model of the marker onto an image plane.

6. The method of claim 4 , further comprising:

receiving an input image;

determining a maximum cross correlation between a part of the input image and the template; and

using a position of the part of the input image to determine a position of the marker.

7. The method of claim 1 , wherein the projection images comprise two-dimensional cone beam computed tomography (CT) projections.

8. The method of claim 1 , wherein the three-dimensional model has an irregular configuration.

9. An image processing method that involves the three-dimensional model of claim 1 , comprising:

obtaining an image that includes an image artifact; and

using the three-dimensional model to remove the image artifact.

10. The image processing of claim 9 , wherein the using the three-dimensional model comprises replacing a part of the image that includes the image artifact with the three-dimensional model or with a projection of the three-dimensional model.

11. A method of determining a model of a marker, comprising:

obtaining projection images, each of the projection images having an image of a marker that indicates a shape of the marker;

determining binary images of the marker for respective ones of the projection images; and

constructing, using a processing unit, a three-dimensional model of the marker using the binary images, the three-dimensional model comprising a set of voxels in a three-dimensional space that collectively indicates three-dimensional shape of the marker;

wherein the act of constructing the three-dimensional model of the marker comprises:

determining a first set of voxels that corresponds with a first one of the binary images of the marker, the first set of voxels aligned with rays projected from the first one of the binary images;

determining a second set of voxels that corresponds with a second one of the binary images of the marker, the second set of voxels aligned with rays projected from the second one of the binary images; and

using intersection of the first and second sets of voxels to determine the set of voxels for the three-dimensional model.

12. A system for determining a model of a marker, comprising:

a non-transitory medium storing projection images, each of the projection images having an image of a marker that indicates a shape of the marker, the marker being a physical object implanted inside a patient; and

a processing unit configured for

determining binary images of the marker for respective ones of the projection images, and

constructing a three-dimensional model of the marker using the binary images, the three-dimensional model comprising a set of voxels in a three-dimensional space that collectively indicates an actual three-dimensional shape of the marker.

13. The system of claim 12 , wherein the set of voxels for the three-dimensional model of the marker comprises a first voxel and a second voxel, wherein the first voxel has a first value associated therewith to indicate a first degree of certainty that the first voxel belongs to the three-dimensional model, and the second voxel has a second value associated therewith to indicate a second degree of certainty that the second voxel belongs to the three-dimensional model, the first value and the second value being different.

14. The system of claim 12 , wherein the processing unit is configured for obtaining the projection images by:

receiving a set of reference images; and

selecting a subset of the reference images as the projection images;

wherein two of the reference images in the selected subset have an angular separation that is at least 15 degrees; and

wherein each of the two of the reference images in the selected subset has a marker contrast that is higher than a prescribed value.

15. A system for determining marker position, comprising the processing unit of claim 12 , wherein the processing unit is further configured for creating a template using the three-dimensional model of the marker.

16. The system of claim 15 , wherein the processing unit is configured to create the template by projecting the three-dimensional model of the marker onto an image plane.

17. The system of claim 15 , wherein the processing unit is further configured for:

receiving an input image;

determining a maximum cross correlation between a part of the input image and the template; and

using a position of the part of the input image to determine a position of the marker.

18. The system of claim 12 , wherein the projection images comprise two-dimensional cone beam computed tomography (CT) projections.

19. The system of claim 12 , wherein the three-dimensional model has an irregular configuration.

20. A system for performing an image processing method that involves the three-dimensional model of claim 12 , comprising the processing unit, wherein the processing unit is further configured for:

obtaining an image that includes an image artifact; and

using the three-dimensional model to remove the image artifact.

21. The system of claim 20 , wherein the processing unit is configured for replacing a part of the image that includes the image artifact with the three-dimensional model or with a projection of the three-dimensional model.

22. A system for determining a model of a marker, comprising:

a non-transitory medium storing projection images, each of the projection images having an image of a marker that indicates a shape of the marker; and

a processing unit configured for

determining binary images of the marker for respective ones of the projection images, and

constructing a three-dimensional model of the marker using the binary images, the three-dimensional model comprising a set of voxels in a three-dimensional space that collectively indicates a three-dimensional shape of the marker;

wherein the processing unit is configured for constructing the three-dimensional model of the marker by:

determining a first set of voxels that corresponds with the a first one of the binary images of the marker, the first set of voxels aligned with rays projected from the first one of the binary images;

determining a second set of voxels that corresponds with a second one of the binary images of the marker, the second set of voxels aligned with rays projected from the second one of the binary images; and

using intersection of the first and second sets of voxels to determine the set of voxels for the three-dimensional model.

23. A computer program product having a non-transitory medium storing a set of instructions, an execution of which will cause a process to be performed, the process comprising:

obtaining projection images, each of the projection images having an image of a marker that indicates a shape of the marker, the marker being a physical object implanted inside a patient;

determining binary images of the marker for respective ones of the projection images; and

constructing a three-dimensional model of the marker using the binary images, the three-dimensional model comprising a set of voxels in a three-dimensional space that collectively indicates an actual three-dimensional shape of the marker.

24. The computer program product of claim 23 , wherein the set of voxels for the three-dimensional model of the marker comprises a first voxel and a second voxel, wherein the first voxel has a first value associated therewith to indicate a first degree of certainty that the first voxel belongs to the three-dimensional model, and the second voxel has a second value associated therewith to indicate a second degree of certainty that the second voxel belongs to the three-dimensional model, the first value and the second value being different.

25. The computer program product of claim 23 , wherein the obtaining the projection images comprises:

receiving a set of reference images; and

selecting a subset of the reference images as the projection images;

wherein two of the reference images in the selected subset have an angular separation that is at least 15 degrees; and

wherein each of the two of the reference images in the selected subset has a marker contrast that is higher than a prescribed value.

26. The computer program product of claim 23 , wherein the process further comprises creating a template using the three-dimensional model of the marker.

27. The computer program product of claim 26 , wherein the template is created by projecting the three-dimensional model of the marker onto an image plane.

28. The computer program product of claim 26 , wherein the process further comprises:

receiving an input image;

determining a maximum cross correlation between a part of the input image and the template; and

using a position of the part of the input image to determine a position of the marker.

29. The computer program product of claim 23 , wherein the projection images comprise two-dimensional cone beam computed tomography (CT) projections.

30. The computer program product of claim 23 , wherein the three-dimensional model has an irregular configuration.

31. The computer program product of claim 23 , wherein the process further comprises:

obtaining an image that includes an image artifact; and

using the three-dimensional model to remove the image artifact.

32. The computer program product of claim 31 , wherein the using the three-dimensional model comprises replacing a part of the image that includes the image artifact with the three-dimensional model or with a projection of the three-dimensional model.

33. A computer program product having a non-transitory medium storing a set of instructions, an execution of which will cause a process to be performed, the process comprising:

obtaining projection images, each of the projection images having an image of a marker that indicates a shape of the marker;

determining binary images of the marker for respective ones of the projection images; and

constructing a three-dimensional model of the marker using the binary images, the three-dimensional model comprising a set of voxels in a three-dimensional space that collectively indicates a three-dimensional shape of the marker;

wherein the constructing the three-dimensional model of the marker comprises:

determining a first set of voxels that corresponds with the a first one of the binary images of the marker, the first set of voxels aligned with rays projected from the first one of the binary images;

determining a second set of voxels that corresponds with a second one of the binary images of the marker, the second set of voxels aligned with rays projected from the second one of the binary images; and

using intersection of the first and second sets of voxels to determine the set of voxels for the three-dimensional model.