CGAL users discussion list

[Thorben Kröger <>]

On Tuesday 13 October 2009 16:01:38 Marc Pouget wrote:
> At first sight i would say you may get better results by adjusting
> the parameters:

> 
> 1. if you set the degree of the fitting to be 1, then you will
> actually get the PCA result which seems good

So I forgot to mention that, sorry.
I set the following parameters via my Matlab interface:
'd_fitting' , 2,
'd_monge', 2,
'nb_points_to_use',32
I verified that indeed these settings arrive in the relevant part of CGAL.
I don't want to set the degree to 1, because curvature estimation might be 
useful. The plane was just a toy problem.

> 
> 2. if you also need curvatures or if the surface seems smooth enough,
> you may want to try to increase the degree to 2. Then it is desirable
> that the point of interest be at the center of the patch of selected
> vertices for the estimation.

Well, I verified for my randomly chosen point that it's roughly at the center 
of all the 32 points that get collected (the blue '*').

> On your example, this symmetry can be improved, instead of selecting
> a given number of vertices (i suppose that's what you did ?) select
> instead a given number of rings. This does not completely solve the
> symmetry problem because of the triangulation artifact but it is
> likely to give a better balance.

I now tried with the options of 4 rings, and I got the result I attach as a 
.png, it looks the same as all surface estimates I've done so far, regardless 
the parameters I chose: The arrows deviated far too much from Vertex to 
Vertex.

With just the PCA smallest eigenvector normals, my 3d vector field looks 
completely homogenous as it should be.

Thanks for your help,

Thorben