CGAL users discussion list

[Pierre Alliez <>]

for parameterization it is better to have a compact disc (large area for small perimeter), and when remeshing using local operators you just need that the underlying patch is parameterized to reach good performances.

size of the disk in terms of #triangles obviously matters for performance issues
shape also matters in the sense that if the patch boundary is closed to a circle you can map it to a circle in parameter space (eg fixed boundary method) with little distortion.


Chaman Singh Verma a écrit :

On Tue, Mar 3, 2009 at 8:12 PM, Pierre Alliez <> wrote:

hi Chaman,

the disks required for the parameterization are...topological disks, ie the input chain
of halfedges must bound a disk.

If  I follow your next email about isotropic remeshing perhaps you are trying to re-implement "Isotropic Remeshing of Surfaces: a Local Parameterization Approach" ? unfortunately I have not put the implementation in CGAL and (I do not have one to send you) you can always check at any point in time during remeshing check that a patch is a disk and in the negative, construct one by region growing while enforcing this property.

pierre

Pierre Alliez
INRIA Sophia Antipolis - Mediterranee 
Project-team GEOMETRICA 
http://www-sop.inria.fr/members/Pierre.Alliez/
Tel: +33 4 92 38 76 77
Fax: +33 4 97 15 53 95
    

Chaman Singh Verma a écrit :

Hello,

In the CGAL Surface Parameterization modules, it is mentioned that it mandatory to provide topological disks
and there are no automatic way to do it and it is still open problem.

Can someone give deeper inside the properties of the disks that are required for parameterizations ?

Also,  I was thinking that many mesh partitioning tools ( such as Metis ) can create lots of surface patches,
aren't those patches good for parameterization ?  What is the opinion of the community on this issue ?

Thanks.

Chaman Singh Verma
Poona, India.

Hello Pierre,

Thanks for your reply, but still I have some doubt. Doesn't metis produces connected patches and what are the
issues because of which this approach in not mentioned in the literature.  Does the shape and size of the disk
matter for good parameterization ?

Thanks.
csv