Skip to Content.
Sympa Menu

cgal-discuss - Re: [cgal-discuss] Question about Nef Polyhedra

Subject: CGAL users discussion list

List archive

Re: [cgal-discuss] Question about Nef Polyhedra


Chronological Thread 
  • From: "Fred Dorosh" <>
  • To:
  • Subject: Re: [cgal-discuss] Question about Nef Polyhedra
  • Date: Thu, 2 Jul 2009 10:24:46 -0400 (EDT)
  • Importance: Normal

> Is it possible to build the Nef-polyhedra from some non-manifold dataset
> like a triangle-soup?
>
> If yes, is there an example of Nef-polyhedron builder class like
> polyhedron
> builder.


Hi Atul,

Since nobody's grabbing this, I'll jump in. There are two answers: 1) yes,
but it's slow and 2) yes, but it's currently difficult.

The standard answer is that for triangle soup, you need to construct your
polyhedron via a nary_union of individual polygon nefs, where each of the
latter is constructed with vertex_cycle_to_nef_3. This is quite slow, both
in practice and asymptotically for large datasets. Your mileage may vary.

As for the seecond answer, it is in fact possible to directly construct
the Nef3 representation using the internal decoration interfaces
(SNC_structure and SM_decorator). If you look at the implementation of
vertex_cycle_to_nef_3, you will get an idea of how this can be done. These
interfaces aren't documented or foolproofed, as they probably weren't
meant to be exposed at this stage.

I've used these interfaces to write my own
non-manifold-triangle-soup-to-nef implementation, under certain
assumptions (weldable, non-self-intersecting input), but it required
certain modifications to the decorator classes themselves. As far as I
know I'm the only one who's done this so far. So the good news is that
it's certainly possible to do.

I'm certain that an official soup-to-nef converter is in the roadmap for
Nef3, if you're able to wait.

Fred






Archive powered by MHonArc 2.6.16.

Top of Page