[vmtk-users] vmtkbifurcationvectors - Computation of bifurcation points
Brought to you by:
davidsteinman,
lucantiga
From: Carlos A. B. <car...@gm...> - 2014-06-05 21:33:19
|
Hi vmtk community, I'm trying to understand the output of the vmtkbifurcationvectors filter. After following the tutorial I successfully compute the bifurcation vectors, but I would like to know a litle more about the output, Suppose I have an Y shaped tree, then the filter gives me a polydata with 3 points (P1,P2,P3), I understand each point is associated to an specific artery (through the GroupId array), and all three belong to the same bifurcation (BifurcationGroupId). I can also know the direction by querying the BifurcationVectorsOrientation array, and angles with the corresponding arrays. What I found difficult to understand is how this 3 points are calculated, because all angles and in/out vectors depend on them, so I try to read he source code first. Please, correct-me if i miss-understood something: First the bifurcation is identify, Then, for each cell i (arterial segment) involved in the bifurcation The orientation (in/out) is determined The subset of cell points at the begin and end of the segment are determined (by the radius) A weighted averaged is used for computing two points (for the beginning and end of the cell) which are called averageTouchingPoint and averageLastPoint Then, depending on the orientation, the bifurcation vector of the cell is computed by subtracting this points, and one of them is used as the bifurcation point Pi. Then, if that is the way of computing P1, P2 and P3, that explains why they usually don't touch the associated centerline, for example for a C shaped branch, the bifurcation point will be inside the C, right? I was thinking also that the bifurcation plane centroid is never used for computing the bifurcation point nor the bifurcation vectors of each cell, correct? For example the bifurcation point could be set to the averageTouchingPoint and the bifurcation vector computed as bifurcationCentroid - averageTouchingPoint. Well Thank you very match for developing VMTK, and i hope you can help me find out the exact way the bifurcation points are computed. Regards Carlos |