Biomedical Engineering Reference
In-Depth Information
(a)
(b)
(c)
(d)
Figure 7.
Comparison between 3D Fast Marching Methods for coronary artery segmenta-
tion and Maximum Intensity Projection: (a) segmentation result for the coronary artery tree
of a mouse heart from 3D micro CT data; (b) Maximum Intensity Projection of the original
data with skeleton; (c) segmentation result for the coronary artery tree of a sheep heart from
3D CT data; (d) Maximum Intensity Projection of the original data with segmentation.
descent can be implemented by point-by-point stepping to the neighbor with the
smallest T-value. The smoothness of the path extracted this way is restricted by the
pixel resolution and might be zigzagged [46]. By using distance information,
the centerline could be easily extracted in binary images, which provides the skele-
ton for the segmentation result. By using a special function of image intensity as
the cost, the method will help to find the path of the vessel in an angiographic
image directly.
