Creating a vascular tree model
A method for vascular modeling is disclosed. The method, in some embodiments, comprises receiving a plurality of 2-D angiographic images of a portion of a vasculature of a subject, and processing the images to automatically detect 2-D features, for example, paths along vascular extents, which are projected into 3-D to determine homologous features among blood vessels. In some embodiments, projection and/or image registration is iteratively altered to improve feature position matching. Based on 3-D vascular extents and their registration to 2-D images, additional features such as vascular width are optionally determined and added to the model.
1. A method for construction of a vascular tree model comprising:
receiving by image processing circuitry a plurality of 2-D angiographic images showing blood vessel segments of a portion of a vasculature of a subject;
extracting automatically by the image processing circuitry, from each of said plurality of 2-D angiographic images, an image feature set comprising 2-D feature positions of said blood vessel segments; and
identifying homologous vascular branches among the 2-D angiographic images using said blood vessel segments, wherein the identifying comprises:
adjusting automatically by the image processing circuitry said 2-D feature positions to reduce relative position error in a common 3-D coordinate system to which each said image feature set is back-projectable, and
then associating automatically by the image processing circuitry image features of the image feature sets as representing homologous vascular branches, based on their projection from the planes of the 2-D angiographic images to common blood vessel segment regions in said common 3-D coordinate system.
2. The method of claim 1 , wherein the image feature set which is extracted comprises a centerline data set including 2-D centerline positions ordered along said blood vessel segments.
3. The method of claim 2 , wherein said adjusting and associating are performed with elements of said centerline data set.
4. The method of claim 2 , wherein said adjusting comprises registration of said 2-D images in 3-D space according to parameters which bring said 2-D centerline positions into closer correspondence among their 3-D back-projections.
5. The method of claim 2 , comprising estimating a metric of radial vascular width based on values of at least one of said plurality of 2-D angiographic images along lines perpendicular to said ordered 2-D centerline positions.
6. The method of claim 2 , wherein said adjusting comprises adjusting of said 2-D feature positions based on projection of said common 3-D coordinate system into the 2-D plane of at least one of said plurality of 2-D angiographic images.
7. The method of claim 1 , wherein information required for said associating automatically of image features is entirely provided before review of said images by a human operator.
8. The method of claim 1 , further comprising defining a surface corresponding to a shape of the heart of the subject, and using said surface as a constraint for said associating of said image features.
9. The method of claim 1 , wherein said images are acquired upon injection of a contrast agent to said vasculature, and further comprising:
determining temporal characteristics of the movement of said contrast agent through said vasculature; and
constraining said 2-D feature positions based on said temporal characteristics.
10. The method of claim 1 , wherein said portion of said vasculature comprises coronary arteries.
11. The method of claim 1 , further comprising:
inspecting projections from the blood vessel segment regions associated with the image features to the planes of 2-D angiographic images to identify image portions representing blood vessel segment regions for which a physical characteristic is to be calculated;
selecting automatically by image processing circuitry a portion of at least one of said plurality of 2-D angiographic images, the portion comprising a representation of said associated image features identified by the inspecting; and
calculating a physical characteristic of at least one of the blood vessel segments, based on the selected image portion.
12. A computer software product, comprising a computer-readable storage medium in which program instructions are stored, which instructions, when read by a computer, cause the computer to receive a plurality of 2-D angiographic images of a portion of a vasculature and execute the method according to claim 1 .
13. A method for construction of a vascular tree model comprising:
receiving by image processing circuitry a plurality of 2-D angiographic images showing blood vessel segments of a portion of a vasculature of a subject;
extracting automatically by the image processing circuitry, from each of said plurality of 2-D angiographic images, an image feature set comprising 2-D feature positions of said blood vessel segments;
adjusting automatically by the image processing circuitry said 2-D feature positions to reduce relative position error in a common 3-D coordinate system to which each said image feature set is back-projectable;
wherein the adjusting comprises iteratively determining a consensus centerline in the common 3-D coordinate system for the feature sets from the plurality of 2-D angiographic images, and wherein the relative position error is calculated based on differences in positions of the projection of the consensus centerline to coordinate systems of the plurality of 2-D angiographic images from positions of the respective image feature sets of the plurality of 2-D angiographic images; and
associating automatically by the image processing circuitry image features of the image feature sets to one another based on their projection from the planes of the 2-D angiographic images to common blood vessel segment regions in said common 3-D coordinate system.
14. The method of claim 13 , wherein the image feature set comprises a centerline data set including 2-D centerline positions ordered along said blood vessel segments.
15. The method of claim 14 , wherein said adjusting and associating are performed with elements of said centerline data set.
16. The method of claim 14 , wherein said adjusting comprises registration of said 2-D images in 3-D space according to parameters which bring said 2-D centerline positions into closer correspondence among their 3-D back-projections.
17. The method of claim 14 , comprising estimating a metric of radial vascular width based on values of at least one of said plurality of 2-D angiographic images along lines perpendicular to said ordered 2-D centerline positions.
18. The method of claim 14 , wherein said adjusting comprises adjusting of said 2-D feature positions based on projection of said common 3-D coordinate system into the 2-D plane of at least one of said plurality of 2-D angiographic images.
19. The method of claim 13 , wherein information for said associating automatically of image features is provided before review of said 2-D angiographic images by a human operator.
20. The method of claim 13 , further comprising defining a surface corresponding to a shape of the heart of the subject, and using said surface as a constraint for said associating of said image features.
21. The method of claim 13 , wherein said images are acquired upon injection of a contrast agent to said vasculature, and further comprising:
determining temporal characteristics of the movement of said contrast agent through said vasculature; and
constraining said 2-D feature positions based on said temporal characteristics.
22. The method of claim 13 , wherein said portion of said vasculature comprises coronary arteries.
23. The method of claim 13 , further comprising:
inspecting projections from the blood vessel segment regions associated with the image features to the planes of 2-D angiographic images to identify image portions representing blood vessel segment regions for which a physical characteristic is to be calculated;
selecting automatically by image processing circuitry a portion of at least one of said plurality of 2-D angiographic images, the portion comprising a representation of said associated image features identified by the inspecting; and
calculating a physical characteristic of at least one of the blood vessel segments, based on the selected image portion.