CGAL users discussion list

[Andreas Fabri <>]

On 14/03/2012 11:24, Zohar wrote:
> There are two options, using c2t3 or c3t3. C2t3 is currently under
> development and buggy. C3t3 works even better than Graphite (on some small
> samples that I tried), but still I'm waiting to solve:
>
> http://cgal-discuss.949826.n4.nabble.com/Convert-C3t3-to-Polyhedron-td4467182.html
>
> to get the output. I was wondering what's the difference between the two
> mesh generation algorithms. They both seem to implement [BO05], which uses
> 3D restricted delaunay. Can I expect in the future from both to generate the
> same output mesh?
>
> Next I implemented my own small idea, which I described on the second post
> on the link above. Namely to represent the polyhedron as an implicit
> function of inside / outside, and feed it to c2t3. Laurent condition proved
> to be more robust than mine. My condition generated far away islands even
> with Epic kernel (??). This simple remeshing proved to be more robust
> (again, on some small tests that I've made) than the previous mentioned
> algorithms.
> I found that for the AABB Simple_cartesian<double>  is enough.
> Any ideas for improving the performance or the inside condition?

If you do many tests, precompute inside/outside for points on
a grid (slightly perturb the points). Then don't shoot a ray
but make a segment query to the closest grid point.

Short segments traverse less AABB tree nodes,

andreas

> --
> View this message in context: 
> http://cgal-discuss.949826.n4.nabble.com/Surface-remeshing-using-mesh-generation-tp4471479p4471479.html
> Sent from the cgal-discuss mailing list archive at Nabble.com.


--
Andreas Fabri, PhD
Chief Officer, GeometryFactory
Editor, The CGAL Project

phone: +33.492.954.912    skype: andreas.fabri