Modeling cerebral aneurysms in medical images
View Patent ↗Methods and systems for modeling cerebral aneurysm and their incoming and outgoing vessels from 3D image data are disclosed. Aneurysms and vessels are segmented from their background using a graph-cuts method. Begin and end of vessels are determined. Construction of a centerline of the incoming and outgoing vessels using a measure of vesselness in calculating a minimum cost path in a graph with nodes being representation of pixels is also disclosed. Vessel surface models are constructed from sub-voxel cross-sectional segmentation. The interpolation of vessels inside an aneurysm based on smooth continuity is disclosed. Selection of endo-vascular stents based on interpolation results is also provided.
1. A method for modeling a cerebral vessel with a beginning and an end with an aneurysm from image data having a plurality of image pixels, comprising:
a processor segmenting the aneurysm and cerebral vessel from a background into a segmentation mask;
the processor determining the beginning and the end of the cerebral vessel;
the processor extracting a local centerline between the beginning and end of the cerebral vessel up to the aneurysm by applying a measure of vesselness based on multi-scale cross-sectional vessel modeling; and
the processor constructing a 3D cerebral vessel surface model up to the aneurysm
wherein the extracting of the centerline includes the steps of:
creating a graph of nodes and edges wherein nodes represent pixels in the plurality of pixels and assigning a cost to an edge, the cost of an edge being determined as a measure of vesselness calculated orthogonal to the edge; and
determining the local center axis by calculating a minimal-cost path in the graph;
wherein the measure of vesselness VM(x,y,z) for a pixel at coordinate (x,y,z) is related to a fit measure f i along a ray for a number of N rays according to an expression
VM
(
x
,
y
,
z
)
=
1
∑
i
=
1
N
f
i
;
and
wherein a fit measure f i along a ray is a difference between a measured intensity profile I and a vessel model V according to an expression
f
i
=
arg
min
R
,
σ
(
γ
1
∑
x
=
0
R
V
i
(
x
,
R
,
σ
)
-
I
i
(
x
)
2
f
(
x
)
+
γ
2
∑
x
=
R
x
B
V
i
(
x
,
R
,
σ
)
-
I
i
2
+
γ
3
∑
x
=
x
B
x
E
V
i
(
x
,
R
,
σ
)
-
I
i
(
x
)
2
g
(
x
)
)
.
2. The method as claimed in claim 1 , wherein the aneurysm and cerebral vessel are segmented from the background into a segmentation mask by comprising the steps:
limiting the segmentation process to image data within a bounding box;
placing an initial seed inside the aneurysm and parental vessels; and
applying graph-cuts optimization originating from the seed to create a segmentation mask for the aneurysm and cerebral vessel.
3. The method as claimed in claim 1 , wherein constructing the 3D cerebral vessel surface model up to the aneurysm includes the steps of:
constructing a plurality of cerebral vessel sub-voxel 2D cross-sectional contours at a plurality of locations of the local center axis of the cerebral vessel; and
constructing a 3D vessel surface model from the plurality of 2D contours.
4. The method as claimed in claim 2 , wherein the beginning and the end of the cerebral vessel are determined from an intersection of the segmentation mask and the bounding box.
5. The method as claimed in claim 2 , further comprising constructing an approximate aneurysm mask by applying a constrained distance transform starting from aneurysm medialness points.
6. The method as claimed in claim 1 , further comprising:
completing the centerline through the aneurysm by using interpolation;
completing the 3D vessel surface through the aneurysm; and
modeling the aneurysm and parental vessels as separate objects in an image.
7. The method as claimed in claim 6 , wherein the interpolation is a bi-spline interpolation between two end points of vessels.
8. The method as claimed in claim 6 , further comprising selecting an endo-vascular stent based on the interpolation.
9. A system for modeling a cerebral vessel with a beginning and an end with an aneurysm from image data having a plurality of image pixels, comprising:
a processor;
software operable on the processor to:
segmenting the aneurysm and cerebral vessel from a background into a segmentation mask;
determining the beginning and the end of the cerebral vessel;
extracting a local centerline between the beginning and end of the cerebral vessel up to the aneurysm by applying a measure of vesselness based on multi-scale cross-sectional vessel modeling; and
constructing a 3D cerebral vessel surface model up to the aneurysm;
wherein the extracting of the centerline includes the steps of:
creating a graph of nodes and edges wherein nodes represent pixels in the plurality of pixels and assigning a cost to an edge, the cost of an edge being determined as a measure of vesselness calculated orthogonal to the edge; and
determining the local center axis by calculating a minimal-cost path in the graph;
wherein the measure of vesselness VM(x,v,z) for a pixel at coordinate (x,y,z) is related to a fit measure f i along a ray for a number of N rays according to an expression
VM
(
x
,
y
,
z
)
=
1
∑
i
=
1
N
f
i
,
and
wherein a fit measure f i along a ray is a difference between a measured intensity profile I and a vessel model V according to an expression
f
i
=
arg
min
R
,
σ
(
γ
1
∑
x
=
0
R
V
i
(
x
,
R
,
σ
)
-
I
i
(
x
)
2
f
(
x
)
+
γ
2
∑
x
=
R
x
B
V
i
(
x
,
R
,
σ
)
-
I
i
2
+
γ
3
∑
x
=
x
B
x
E
V
i
(
x
,
R
,
σ
)
-
I
i
(
x
)
2
g
(
x
)
)
.
10. The system as claimed in claim 9 , wherein the aneurysm and cerebral vessel are segmented from the background into a segmentation mask, comprising the steps:
limiting the segmentation process to image data within a bounding box;
placing an initial seed inside the aneurysm and parental vessels; and
applying graph-cuts optimization originating from the seed to create a segmentation mask for the aneurysm and cerebral vessel.
11. The system as claimed in claim 9 , wherein constructing the 3D cerebral vessel surface model up to the aneurysm includes the steps of:
constructing a plurality of cerebral vessel sub-voxel 2D cross-sectional contours at a plurality of locations of the local center axis of the cerebral vessel; and
constructing a 3D vessel surface model from the plurality of 2D contours.
12. The system as claimed in claim 10 , wherein the beginning and the end of the cerebral vessel are determined from an intersection of the segmentation mask and the bounding box.
13. The system as claimed in claim 10 , further comprising constructing an approximate aneurysm mask by applying a constrained distance transform starting from aneurysm medialness points.
14. The system as claimed in claim 9 , further comprising:
completing the centerline through the aneurysm by using interpolation;
completing the 3D vessel surface through the aneurysm; and
modeling the aneurysm and parental vessels as separate objects in an image.
15. The method as claimed in claim 14 , wherein the interpolation is a bi-spline interpolation between two end points of vessels.
16. The system as claimed in claim 14 , further comprising selecting an endo-vascular stent based on the interpolation.