[Insight-users] Registration + Statistical Shape Models

[Marcelo Elias de Oliveira]

Dear Luis and ITK friends,
Congratulation on your work and sorry for the newbie question.
My target is to build a statistical shape model based on a training set (DICOM images). Consequently, correspondences between landmarks (i.e., points in the reference and moving surfaces) should be established before applying Procrustes. 
In order to achieve the point-to-point correspondences, I am performing rigid + affine + non-rigid registration between the fixed and moving images, as shown on “DeformableRegistration15.cxx“. 
Currently, I am having problems in order to deform my moving surface and preserve the same structure of the reference surface (i.e., number of points, etc). Is it possible to preserve the same structure of the fixed image using the itkResampleImageFilter + vtkMarchingCubes? 
I tried to use other procedures based on BinaryMask3DMeshSource and DeformableMesh3DFilter filters, without success.
Any comment will be greatly appreciated!
Thank you for your attention!
Regards,
Marcelo

P.S.: I am trying to register complex geometries such as vertebral bodies. Thus, any other information regarding preprocessing of images, and-or suggestion about other methods such as ICP will be also greatly appreciated.

-- 
GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT!
Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01