CGAL users discussion list

["Satyajit Sarangi" <>]

Hi Monique,
           I did try this approach you suggested before but it still didn't solve my problem. Essentially my problem boils down to extracting a convex hull from the delaunay triangulation. I don't know whether you got the previous screenshot of what exactly my problem was, when doing this hence I am sending the link again. 

http://www.eng.buffalo.edu/~ssarangi/Triangulation.bmp

The problem is that of the triangles which are pointing inwards (in the left image) and also the triangles on the surface somehow look a bit odd suggesting that all the triangles on the hull haven't been included. Laurent has suggested the same method of getting the infinite vertex index and then extracting the triangle from that. But that also didn't work.
Thanks,
Satyajit

On 8/8/07, <> wrote:

Hi Satyajit,

If I understand well, using the following method of Triangulation_3
should solve your problem.

Triangle          t.triangle ( const Cell_handle c, int i)
                Returns the triangle formed by the three vertices of facet (c,i). The
triangle is oriented so that its normal points to the inside of cell c.
Precondition:   t.dimension() greater or equal 2 and i in {0,1,2,3} in
dimension 3, i = 3 in dimension 2, and the facet is finite.

use it with i = the index of the infinite vertex in your infinite cell.

In this method, we take care of whether i is even or odd.
The user manual explains this orientation issue.

best
        Monique Teillaud

Satyajit Sarangi wrote:

> I don't understand why the triangles somehow point to some interior
> point of the point cloud. I mean the interior triangles. Essentially it
> should be the convex hull but thats not what I am getting. I am only
> including the triangles which are in my "TriangleList" for rendering.
> Please see the screenshot attached. Laurent, I will definitely try out
> the convex hull approach but I am trying to see if one of my filtration
> algorithm works on the delaunay and hence need it. I know its an
> overkill but right now I have to see both the approaches.
> Thanks and Regards,
> Satyajit
--
You are currently subscribed to cgal-discuss.
To unsubscribe or access the archives, go to
https://lists-sop.inria.fr/wws/info/cgal-discuss