CGAL users discussion list

[Andreas Fabri <>]

Hello,

Let me give you a quick incomplete answer. Maybe the authors of the
package will follow up.

The result seems to be correct modulo the sign of the vector.
And there I am not sure how easy it is in general to make it
consistent. It needs probably one place to start and has to
propagate information around in order to make vectors show
in similar enough directions.

best regards,

andreas

Oleksiy Busaryev wrote:
> Hello guys,
>  
> I have the following problem.
>  
> I want to use CGAL for computing principal directions of curvature for 
> each vertex of a triangular mesh.
>  
> I tried to use code from Jet_fitting_3 example to do this - but even 
> if I analyze large number of rings around each vertex,
> principal curvature directions do not look consistent: for adjacent 
> vertices (non-umbilic) sometimes they are opposite.
>  
> I attached a simple example: triangular mesh of a cylinder with 
> principal directions shown
> (3 rings was computed around each vertex for Monge fitting).
>  
> You can see that directions of maximum curvature do not go around the 
> cylinder in a consistent manner (as I expected).
>  
> What can be the reason for this?
>  
> Thank you!
>
> --
> Best regards,
> Oleksiy Busaryev.
>
> ------------------------------------------------------------------------