Skip to Content.
Sympa Menu

cgal-discuss - Re: [cgal-discuss] How General is Triangulation_3?

Subject: CGAL users discussion list

List archive

Re: [cgal-discuss] How General is Triangulation_3?


Chronological Thread 
  • From: Olivier Devillers <>
  • To:
  • Subject: Re: [cgal-discuss] How General is Triangulation_3?
  • Date: Tue, 26 Aug 2008 09:29:01 +0200


a écrit :
I am interested in knowing how general the Triangulation_3 class is. It
requires that the triangulation be a triangulation of a 4-sphere; the basic
idea is that the triangulation is a finite triangulation of some convex body
together with a point at infinity.

In some cases it would be possible to extend a tetrahedral mesh of a
nonconvex body to a mesh of the convex hull of the mesh vertices, but I
believe it is not always possible to do this without adding new vertices.
If the boundary of your object is non convex, you may link all facets of the boundary to the point at infinity
and then overload the is_valid test of a facet such that it does not check convexity.

For example, consider a tetrahedral mesh with a hole in the shape of a
Schonhardt polyhedron.
If you have holes, then you can add one "virtual point"per hole, such a point play the same role as point at infinity.
You need to be able to distinguish between "usual tetrahedra" and "virtual tetrahedra" to overload the is_valid test

a virtual tetrahedron is a tetrahedron with a virtual vertex (or the vertex at infinity).
vertices can have an extra field "virtual" or be stored in a map
(in current implementation, the only virtual vertex = vertex at infinity is stored in the triangulation object to be compared to any vertex)





Archive powered by MHonArc 2.6.16.

Top of Page