CGAL users discussion list

["Sebastien Loriot (GeometryFactory)" <>]

You can try using the functions keep/remove_connected_components.

http://doc.cgal.org/latest/Polygon_mesh_processing/group__keep__connected__components__grp.html

http://doc.cgal.org/latest/Polygon_mesh_processing/group__keep__connected__components__grp.html

The problem might be that the output is non-manifold (two patches 
sharing only one vertex).

Sebastien.


On 06/02/2016 03:01 PM, Sukhraj Singh wrote:
> Hello Sebastien,
>
> I tried the following example to create a polyhedron from a segment in
> an incremental manner.
> http://doc.cgal.org/latest/Polyhedron/Polyhedron_2polyhedron_prog_incr_builder_8cpp-example.html
>
> I followed following procedure (In brief).
> - Iterate the facets
> If facet belongs to i-th segment.
> Push its points to the incremental builder
> Then add vertex to facet.
>
>
> I got following error.
> CGAL::Polyhedron_incremental_builder_3<HDS>::
> lookup_halfedge(): input error: facet 1 shares a halfedge from vertex 1
> to vertex 2 with facet 0.
>
> I am getting the reason what is happening, but the sample provided
> "polyhedron_prog_incr_builder.cpp" is too simplistic.
> Is there any other example where above case is handled. Because I think
> it can happen very often.
> Any hint to resolve the above issue would be appreciated.
>
>
>
> Thanks
>
>
>
> On Mon, May 23, 2016 at 8:38 PM, Sebastien Loriot (GeometryFactory)
> <
>  
> <mailto:>>
>  wrote:
>
>     If it does what you want, then yes.
>
>     Sebastien.
>
>     On 05/23/2016 04:57 PM, Sukhraj Singh wrote:
>
>         Okay,
>
>         Assuming I don't use connected_components().
>         And I directly create Polyhedrons from the segments, i.e.
>
>         - Iterate through polyhedron.
>         - Create polyhedron of a segment using
>         Polyhedron_incremental_builder_3.
>
>         Is this correct way to do ? Or there is some better approach in
>         CGAL.
>
>
>
>
>         On Mon, May 23, 2016 at 1:18 PM, Sebastien Loriot (GeometryFactory)
>         
> <
>  
> <mailto:>
>         
> <mailto:
>  
> <mailto:>>>
>  wrote:
>
>              What I mean is if you need both clusters and segments but
>         want to
>              save on computation time, you can run the segmentation
>         algorithm to
>              extract
>              segments and then clusters using `connected_components()`
>         (instead of
>              another run of segmentation).
>
>              Sebastien.
>
>              On 05/22/2016 02:32 PM, Sukhraj Singh wrote:
>
>                  Thanks for your answer, Sebastien
>                  Do I really need to use connected_components()
>
>                  I already have the "segment_pmap", where the value is the
>                  segment id to
>                  which that particular face belongs.
>                  And one could parse the Polyhedron and collect all the
>         faces having
>                  particular segment ID.(Of-course I need to check, how to
>                  implement the same)
>                  Is my understanding correct ?
>
>
>
>                  Further, quoting from my previous email:
>
>                       Two spatially apart areas( i.e. not having
>                       common edges) of a Mesh may fall into same
>         segment, since
>                  they had
>                       comparable SDF values.
>
>
>                  I am not sure about it, they of-course may fall into same
>                  cluster, but
>                  may not be to the same segment.  In such a case if,
>         faces that don't
>                  share edge, will fall into different segment always,
>         even though
>                  they
>                  were in same cluster, using connected_components() may
>         not be
>                  required ?
>
>
>
>
>                  On Thu, May 12, 2016 at 6:00 PM, Sebastien Loriot
>         (GeometryFactory)
>                  
> <
>  
> <mailto:>
>         
> <mailto:
>  
> <mailto:>>
>                  
> <mailto:
>         
> <mailto:>
>  
> <mailto:
>         
> <mailto:>>>>
>  wrote:
>
>                       On 05/12/2016 02:16 PM, Sukhraj Singh wrote:
>
>                           Salut,
>
>                           The "Triangulated Surface Mesh Segmentation"
>         module
>                  allows one to
>                           segment a surface mesh using the SDF values.
>                           It creates a Property Map which is colorized
>         according
>                  to the
>                           indices,
>                           resulting into segments. Two spatially apart
>         areas(
>                  i.e. not having
>                           common edges) of a Mesh may fall into same
>         segment,
>                  since they had
>                           comparable SDF values.
>
>                           My question is, how can one extract these
>         segments.
>                  Since I want
>                           to do
>                           further processing on a segment separately. Also I
>                  would like
>                           further
>                           divide a segment, into say sub-segments , where a
>                  sub-segment
>                           will have
>                           faces which are not spatially apart, i.e.  a
>         subset
>                  which is having
>                           common edges only.
>
>
>
>                       Set the parameter output_cluster_ids to true when
>         calling the
>                       segmentation function
>
>         
> http://doc.cgal.org/latest/Surface_mesh_segmentation/group__PkgSurfaceSegmentation.html#ga8533ea7b1062ebbabe3bf23fed303f24
>
>                       If you need also the segments, then set the option
>         to false and
>                       use the function connected_components() considering as
>                  constrained
>                       edges the edges between two faces belonging to
>         different
>                       segments.
>
>         
> http://doc.cgal.org/latest/Polygon_mesh_processing/group__keep__connected__components__grp.html#gabe2052b2281e1e628e68d8fea94be2d9
>
>
>                       Sebastien.
>
>
>
>                           Thanks to inform.
>
>
>
>
>                           P.S. I am a newbie.
>                           --
>                           Sukhraj Singh
>
>
>
>                       --
>                       You are currently subscribed to cgal-discuss.
>                       To unsubscribe or access the archives, go to
>         https://sympa.inria.fr/sympa/info/cgal-discuss
>
>
>
>
>
>                  --
>                  Sukhraj Singh
>
>
>
>              --
>              You are currently subscribed to cgal-discuss.
>              To unsubscribe or access the archives, go to
>         https://sympa.inria.fr/sympa/info/cgal-discuss
>
>
>
>
>
>         --
>         Sukhraj Singh
>
>
>
>     --
>     You are currently subscribed to cgal-discuss.
>     To unsubscribe or access the archives, go to
>     https://sympa.inria.fr/sympa/info/cgal-discuss
>
>
>
>
>
> --
> Sukhraj Singh