CGAL users discussion list

[Jane Tournois <>]

Hello,

the indices that appear after each triangle and tetrahedron correspond 
to the domain index for each simplex :
- for a facet : the surface patch index
- for a cell : the subdomain index (and -1 outside the domain)

In a labellized image, the indices correspond to the labels available in 
the image.
In a gray level image, the indices are either -1 or 1 (inside/outside)

For more details, you can have a look at the mesh domain concept :
http://doc.cgal.org/latest/Mesh_3/classMeshDomain__3.html

and at the file where the output function is coded :
Mesh_3\include\CGAL\IO\File_medit.h

Best,

Jane.

--
Jane Tournois, PhD
R&D Engineer at GeometryFactory
http://www.geometryfactory.com/




Le 27/01/2016 15:27, Pierre Hallot a écrit :
> Hello,
>
> Thanks a lot for your answer, it works fine.
>
> I have one last question, for the "liver.inr" example, if I have a look at 
> the out.mesh file, the last parameters for vertices (and triangles/tetrahedra) is an 
> integer between -1 and 3, which I think represent if they are on the boundary (2) or 
> inside (3) or I guess isolated for the rest.
> Are those numbers accurate?
>
> With my own .inr file (I have attached it to this message in case) they are 
> all at 1.
> I've tried to add "c3t3.set_dimension(ch->vertex((i+3)% 4),2);" and 
> "c3t3.set_dimension(cit->vertex(i),3);" in the code you provided, but this seems to have no effect. 
> (The "c3t3.output_to_medit(medit_file);" call is made afterwards, I've check.)
>
> I need to find which inner vertices are linked together, so simply using the 
> .mesh file would be great, but now that I have the inner vertices I am sure I 
> can find another way.
>
>
> This is indeed as part of my internship and thus for my company, is this an 
> issue? We will comply with the GPL3 licence, if this is a licence issue.
>
> Kind regards,
> Pierre Hallot
> ________________________________________
> From: 
>
>  
> []
>  on behalf of Andreas Fabri 
> []
> Sent: Wednesday, January 27, 2016 1:07 PM
> To: 
>
> Subject: Re: [cgal-discuss] Finding inner vertices in 
> mesh_optimization_lloyd_example
>
> Hello,
>
> I've rewritten the program you posted and I've put it on github:
> https://gist.github.com/afabri/32a7b4082208bd53b78d
>
> I did not really try to understand your program,
> so let me just make some remarks:
>
> Note that when you have a cavity the infinite vertex is
> not a vertex of a cell.
>
> Also never create or copy a Cell, only use Cell_handle
> or Cell_iterator and access data with ->
>
> You cannot fully control the number of inner vertices,
> as the mesh generator will in any case Delaunay tetrahedra.
>
> There are less vertices if you do not set the
> cell_radius_edge_ratio
>
> Best,
>
> Andreas
>
> On 26/01/2016 14:36, Pierre Hallot wrote:
> > Hello,
> >
> > I have a few questions on the mesh_optimization_lloyd_example that can
> > be found in cgal/Mesh_3/examples/Mesh_3/mesh_optimization_lloyd_example.
> > I am currently using the provided example (liver.inr.gz) and just trying
> > to interact with the results.
> >
> > First, I am trying to find the points that are inside the created mesh,
> > and to differentiate them from those that are on the boundary.
> >   From what I've read, I can find them based on the fact that a vertex on
> > the boundary will have one adjacent vertex which is infinite.
> > I thus made the following code that I added at the end of the file
> > without modifying anything else:
> >
> > Tr t = c3t3.triangulation();
> >
> > // sets so that the same points are not added twice
> > std::set<C3t3::Vertex_handle> innerVertices;
> > std::set<C3t3::Vertex_handle> outerVertices;
> >
> > for (C3t3::Cells_in_complex_iterator cell =
> > c3t3.cells_in_complex_begin(), end = c3t3.cells_in_complex_end(); cell
> > != end; ++cell)
> > {
> >       const Tr::Cell c(*cell);
> >
> >       // each tetrahedra has 4 vertices
> >       for (int i = 0; i < 4; ++i)
> >       {
> >            // the vertices connected to the one we are checking
> >            std::vector<C3t3::Vertex_handle> adjacent_vertices;
> >            t.adjacent_vertices(c.vertex(i),
> > std::back_inserter(adjacent_vertices));
> >            bool onBoundary = false;
> >            int counter = 0;
> >            for (std::vector<C3t3::Vertex_handle>::iterator it =
> > adjacent_vertices.begin(); it != adjacent_vertices.end(); ++it)
> >            {
> >                if (t.is_infinite(adjacent_vertices.at(counter)))
> >               {
> >                    onBoundary = true;
> >                    break;
> >               }
> >                ++counter;
> >            }
> >
> >            if (!onBoundary)
> >            {
> >                innerVertices.insert(c.vertex(i));
> >            }
> >            else
> >            {
> >               outerVertices.insert(c.vertex(i));
> >            }
> >       }
> > }
> >
> > std::cout << "CountInner: " << innerVertices.size() << std::endl;
> > std::cout << "CountOuter: " << outerVertices.size() << std::endl;
> >
> > However, no vertex seems to be infinite and thus every point is
> > considered as inside the complex. How is it possible?
> >
> >
> > Related to this, the File_medit.h outputs a dimension in the .mesh file
> > (the fourth parameter).
> > I think I read somewhere that 3 means it is inside the complex while 2
> > is on the boundary.
> > That's exactly what I need, however for the liver, there are about 2000
> > vertices (out of a total of 3361) with dimension 3 and less than a
> > thousand with dimension 2, the rest has a dimension of 1, 0 or -1.
> > I would have expected to have much more vertices on the boundary than
> > inside the complex however, is what I am saying correct?
> > Moreover, if I replace the liver.inr.gz file by my own .inr file,
> > everything has a dimension of 1, which makes this method useless.
> >
> >
> > Lastly, is it possible to fix the number of vertices inside the complex
> > that will be created during the make_mesh_3 call? For instance, I would
> > like to have only 256 vertices inside. (I don't care about the number on
> > the boundary.)
> >
> > Kind regards,
> > Pierre Hallot
> --
> Andreas Fabri, PhD
> Chief Officer, GeometryFactory
> Editor, The CGAL Project
>
> phone: +33.492.954.912    skype: andreas.fabri
>
> --
> You are currently subscribed to cgal-discuss.
> To unsubscribe or access the archives, go to
> https://sympa.inria.fr/sympa/info/cgal-discuss