CGAL users discussion list

[Sebastien Loriot <>]

If you want to get the same output as Bruno (but with the correct 
orientation ;) ), you can use the upcoming autorefine function like in 
this example:

https://github.com/sloriot/cgal/blob/PMP-BF_autorefine/Polygon_mesh_processing/examples/Polygon_mesh_processing/soup_autorefinement.cpp

The code is robust and has been tested on the full 10k models of
Thingi10k dataset without any issue.

However, this is not a magic tool. This will only refine the 
self-intersecting triangles. Luckily in your case, that was the only issue.
Extra filtering of triangles will need some extra care, as in general
such a technique creates non-manifold meshes.

Best,

Sebastien.

On 7/19/23 13:11, Giles Puckett wrote:
> Attached here as an OFF file.
>
> G.
>
> On 19/07/2023 5:22 pm, Sebastien Loriot ( via 
> cgal-discuss Mailing List) wrote:
> > Could you share the mesh?
> >
> > Best,
> >
> > Sebastien.
> >
> > On 7/19/23 09:15, Giles Puckett wrote:
> > > Thanks Sebastien,
> > >
> > > Unfortunately remove_self_intersections returns false (and the mesh 
> > > is unchanged, still returning the same self intersections as before). 
> > > This is with CGAL 5.5.1, if anything has changed since.
> > >
> > > G.
> > >
> > > On 19/07/2023 3:54 pm, Sebastien Loriot ( via 
> > > cgal-discuss Mailing List) wrote:
> > > > Hello,
> > > >
> > > > you can try 
> > > > CGAL::Polygon_mesh_processing::experimental::remove_self_intersections()
> > > > provided by CGAL/Polygon_mesh_processing/repair_self_intersections.h.
> > > >
> > > > If you only have local self-intersection, it usually works very well.
> > > >
> > > > For listing all intersecting triangles use `self_intersection()`
> > > > https://doc.cgal.org/latest/Polygon_mesh_processing/group__PMP__intersection__grp.html#ga6c5bed399d14ddbba820ce50d3145712
> > > >
> > > > HTH,
> > > >
> > > > Sebastien.
> > > >
> > > > On 7/19/23 07:04, Giles Puckett wrote:
> > > > > Hello all,
> > > > > I have a triangle mesh that contains self-intersections 
> > > > > (PMP::does_self_intersect returns true) but it is otherwise 
> > > > > manifold. Is there an easy way to repair these meshes?
> > > > >
> > > > > If not, would it work to enumerate all the self-intersecting 
> > > > > triangles, delete them from the mesh, and fill the holes left 
> > > > > behind? Is there a good way to do this, or would some other method 
> > > > > work better?
> > > > >
> > > > > I suspect that the affected area is very small, so gross physical 
> > > > > distortion is unlikely to result.