CGAL users discussion list

[Nico Kruithof <>]

Hi Daniel,

I did the same for the skin surface package, where I have the vertices
and tetrahedra of the triangulation and want to reconstruct the neighbor
information. The class is not documented and there is no guarantee that
it works for your cases as well. 
You could have a look at that. It's in the file
CGAL/Triangulation_incremental_builder_3.h

You give the triangulation in the constructor, then you start building
the triangulation with "begin_triangulation(int dim)". I don't remember
whether dim<3 works as I don't use it.
You build the triangulation with the functions: 
  Vertex_handle add_vertex();

  Cell_handle add_cell(Vertex_handle vh0, Vertex_handle vh1,
                       Vertex_handle vh2, Vertex_handle vh3);

Finally, you build the triangulation with
  void end_triangulation()
which also creates the infinite cells and might do some validity
checking.

Let me know whether you find it usefull and whether it could be
improved.

Bests,
Nico

On Tue, 2008-06-03 at 16:13 +0200, 

 wrote:
> Hi at all!
> 
> I'm using the Triangulation_3 data structure for simulating deformable 
> models. The advantage of that data structure is the simple access for 
> information about neighbors. 
> 
> I constructed different models, in my case simple bars, by inserting 
> equidistant points in a Delaunaytriangulation. Using the I/O functionality 
> of the data structures i can load and save these Triangulations. 
> 
> Now i want to simulate other models. I've got the tetrahedral mesh as list 
> of points and tetrahedra referring to the points. I've already spent a long 
> time for an incremental construction of the tetrahedral mesh. But i failed, 
> because of the very complex neighbor information especially of infinite 
> tetrahedrons. 
> 
> Is there any existing algorithm for that kind of problem? Or can anybody 
> give me a hint how to construct the tetrahedral mesh properly? 
> 
> Best regards, 
> Daniel Weber