PSVM involves three fundamental steps: (1) imageprocessing, (2) analysis suitable model generation, and (3)analysis (Fig. 1). First, the scanned images are segmented,or classified into meaningful regions to extract the geom-etry of the volumes of interest. Next, an analysis suit-able solid geometric model is generated from thesegmented images. Typically, solid models of blood ves-sels are constructed by (a) extracting a set of points (acontour) approximating the inner boundary (the lumen-wall interface) of a vessel on an offset stack of segmentedtwo-dimensional (2D) image slices, (b) interpolating thecontour with a curve, (c) lofting a surface through theinterpolated curves, (d) joining the surfaces together toform a network, and (e) meshing, or infilling, the boundedsurfaces to define the volume of the vascular network