Subject: CGAL users discussion list
List archive
- From: Stephane Tayeb <>
- To:
- Subject: Re: [cgal-discuss] 3D mesh generation issues
- Date: Wed, 03 Feb 2010 09:16:29 +0100
Hi,
wrote:
hi!
i have some questions about 3D mesh generation. not programming problems, but
regarding parameters, underlying algorithms, etc.
so far i've had some pretty good results (everything installed and compiles in
visual studio 2008), but there are some limitations/issues.
i'm trying to achieve a SMOOTHLY GRADED mesh from a segmented volumetric
image
(using liver.inr for now). FYI, i've got CGAL all compiled and running in VS2008. examples work! :)
2 questions are numbered as follows:
1. i cannot understand this from the manual: is it required/necessary/better
to
"refine" a coarse mesh in 'steps' or can you just assign facet and cell
parameters from the onset to achieve accurate surface and mesh integrity?
the manual implies (fig. 47.2) that you should start with a course mesh, and
then refine in steps 'from the previous mesh', first by surface accuracy, and
then by resolution. is this necessary to achieve a well configured (i.e.
'good'
tets) mesh??
You don't have to proceed by 'steps'. You should use as parameters those which fulfill your needs directly.
In fig 47.2, all meshes are created from scratch, but you're right the explanations are far from clear. Actually, the idea is to show how parameters change the output. So we just change one parameter from one mesh to the next one (but we always create a new mesh).
Thanks for reporting, this will be fixed.
if so, fine, i just need to know from a programming standpoint.
2. (more important) is there any way to control the mesh density directly. for
example, in 'tetgen' (http://tetgen.berlios.de/) you can control mesh density
via an auxiliary volumetric density map. this is crucial to achieving a well
prescribed mesh.
There is no such thing in the current release.
Nevertheless such a behavior is achievable. You'll have to create your own set of criteria.
For instance, you may write your own sizing criterion for tets, using a function f(Point_3) instead of a fixed size. You could start at this page (http://www.cgal.org/Manual/last/doc_html/cgal_manual/Mesh_3_ref/Concept_MeshCriteria_3.html#Cross_link_anchor_1421) of documentation and have a look at file Mesh_criteria_3.h and in its dependencies.
thanksinadvance,
todd
drexel university
Best,
Stéphane.
--
Stephane Tayeb
Software engineer - INRIA Sophia Antipolis
Geometrica Project-Team
- [cgal-discuss] 3D mesh generation issues, tcdoe, 02/03/2010
- Re: [cgal-discuss] 3D mesh generation issues, Stephane Tayeb, 02/03/2010
- Re: Re: [cgal-discuss] 3D mesh generation issues, tcdoe, 02/23/2010
- Re: [cgal-discuss] 3D mesh generation issues, Stephane Tayeb, 02/03/2010
Archive powered by MHonArc 2.6.16.