[Insight-users] DeformableModelSimplexMesh

[alex. gouaillard]

Dear leila, dear all.

FYI, we are almost ready (one month) to release a new mesh structure in
itk. This mesh structure for discrete surfaces is based on C-GAL and the
so called "Quad-Edge" structure.
The impact on the curent problem you're addressing here is twofold:

1. This structure natively hold primal and dual connectivity information
of the surface, making the actual simplex related classes kind of
obsolete. I would like to volunteer to recode the existing classes
(well, triangle to simplex and simplex to triangle are not necessary
anymore) using the new structure and I was wondering if you would mind
giving me some e-mail assistance if I have any problem? I plan to do
that within the following 30 days.

2. We are planing to port most of the surface processing filters from
vtk to itk to be able to have all the surface processing done in itk,
and only the vizualization in vtk. Has we implemented the euleroperators
in the new structure, this work is made easier. We have already ported
decimate (first version, no topology modification), smoothing, and
implemented an improved version of connectivity filter. More are on the
way. 

cheers,

alex

On Thu, 2006-08-24 at 02:50, Leila Baghdadi wrote:
> Hi Zein,
> 
> I have done some further investigation with this and fixed a few minor
> things which I will commit to cvs soon.
> 
> However, the major issue which I still have not solved is as follows,
> 
> As you know, this application reads a binary a image to be converted to
> a mesh using the vtkMarchingCubes class. This class happens to create
> 4-5 triangles per voxel of image (i.e, a huge number of triangles)
> converting such a mesh to a simplex mesh for deformation will result in
> a disaster unless you have unlimited memory on your system. 
> 
> So To avoid the above I placed the vtkDecimate class right after
> vtkMarchingCubes for the sole purpose of reducing the number of
> triangles by a huge number (i.e, creating a simplex mesh which can
> easily be deformed for many iteration)
> 
> Unfortunately, over the course of VTK going through major modifications
> this class is replaced by vtkDecimatePro which according to my tests
> does not do a good job of reducing the number of Triangles unless
> PreserveTopology is Off at which case it creates holes in the mesh
> (i.e., not usable for converting to simplex mesh and going through the
> deformation)
> 
> I have sent an e-mail to vtk list asking if the old class could be added
> to vtk tree so I could modify my itk code to use the old class.
> Meanwhile, I am going to have to think about this and see if I can
> resolve this myself.
> 
> Please continue using itk,
> 
> Leila 
> 
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users