CGAL users discussion list

[Sebastian Bachmann <>]

Thanks for the hint, this seems to be indeed the problem!
Calling the script now several times reduces the chances that this 
happens, however after a few trials I got to a random placement which 
had again small elements.
I have to read the code a bit to understand how the points are 
constructed, but I think that is do-able!

Thanks for the quick help!

Best,
Sebastian

On 24.05.2023 14:57, Andreas Fabri wrote:
> Have a look at this section
>
> https://doc.cgal.org/latest/Mesh_3/index.html#title30
>
>
> On 5/24/2023 2:56 PM, Andreas Fabri wrote:
> > Hello,
> >
> > Our meshing experts don't work today.   My guess is that it has to do 
> > with the initialization
> > of the mesh generation process.   From a point P inside the volume we 
> > shoot rays and
> > the intersection points of the rays with the surface are inserted. 
> > When the point P
> > is close to the surface you may get two intersection points close on 
> > the surface.
> >
> > Andreas
> >
> >
> > On 5/24/2023 2:51 PM, Sebastian Bachmann ( via 
> > cgal-discuss Mailing List) wrote:
> > > Hi!
> > >
> > > This is an image of a set of 20 and I can see this behaviour on 4 of 
> > > them. On 420.inr it does not happen and the mesh looks good.
> > > The code was compiled against 5.4 (1050401000).
> > >
> > > If I look at the imges, I can not see any voxels standing out at the 
> > > position - at least not more than on other places, where it works fine.
> > > At least from the image, I cannot make out any particular feature 
> > > that would cause that.
> > >
> > > Is there any way to debug this and see what's going on?
> > >
> > > Best,
> > > Sebastian
> > >
> > > On 24.05.2023 14:34, Sebastien Loriot ( via 
> > > cgal-discuss Mailing List) wrote:
> > > > Do you observe this behavior on all images or only a specific one?
> > > > For example, when using 420.inr do you have such an issue?
> > > >
> > > > What could case such an over refinement (provided you are not using 
> > > > master version of CGAL) is if the mesher is trying to respect some
> > > > topological criteria (a very tiny handle for example, but that would 
> > > > be surprising in an image).
> > > >
> > > > Best,
> > > >
> > > > Sebastien.
> > > >
> > > > On 5/24/23 14:14, Sebastian Bachmann ( via cgal-discuss 
> > > > Mailing List) wrote:
> > > > > Hi everyone,
> > > > >
> > > > > I have an image from which I want to create a coarse surface mesh.
> > > > > The code I use for that is based upon the example 
> > > > > mesh_3D_image_with_features.cpp, however it is adopted to then 
> > > > > extract the surface and write an OFF file: 
> > > > > https://gist.github.com/reox/0df1f446f0ff9d46e0688b60e8e858b8
> > > > >
> > > > > As you can see, the Mesh_criteria is set to a really coarse mesh. 
> > > > > The input image (attached) I use has a voxel size of around 0.5mm 
> > > > > and represents a proximal femur.
> > > > > Now, for some reason, I get extremely fine triangle elements in 
> > > > > certain places (see attached image).
> > > > >
> > > > > I find that peculiar, as the facet_distance is already quite large, 
> > > > > and I thought that this might be the main issue here. I remembered, 
> > > > > that also the relative_error_bound in the 
> > > > > create_labeled_image_mesh_domain can have an influence, but 
> > > > > reducing this rather creates a "spiky" mesh than removing those 
> > > > > small elements. (Eventually when it is set to 0.1, the small 
> > > > > elements vanish, but then the vertices are all over the place and 
> > > > > not on the isosurface anymore...)
> > > > >
> > > > > Is there anything I missed here, maybe did I use wrong parameters? 
> > > > > Or is there another way how I could produce such coarse meshes?
> > > > >
> > > > > Best,
> > > > > Sebastian