[vtkusers] compute a geodesic distance

[Rasmus Reinhold Paulsen]

After some digging on the internet I have to revise my statements. It                       
seems that the development on geodesics are somewhat more evolved than I                    
believed and that Fast Marching is not necessarely "The Solution (tm)".                     
                                                                                            
I also located a recent paper dealing with the subject:  

"Fast Exact and Approximate Geodesics on Meshes" V. Surazhsky, T.
Surazhsky, D. Kirsanov, S. J. Gortler and H. Hoppe Proceedings of SIGGRAPH
2005

that can be downloaded here:

http://www.cs.harvard.edu/~sjg/papers.html

VTK implementation anyone?

Cheers,
Rasmus

On Tue, 14 Jun 2005, Steve M. Robbins wrote:

> On Mon, Jun 06, 2005 at 11:36:27PM +0200, Rasmus Reinhold Paulsen wrote:
> > You can do a "pseudo-geodesic" distance using the Dijkstra class available
> > here:
> > 
> > http://www2.imm.dtu.dk/~rrp/VTK/
> 
> An easy improvement on simple Dijkstra is to add extra vertices along
> edges.  See 
> 
> Lanthier et al. 
> Approximating Weighted Shortest Paths on Polyhedral Surfaces
> Algorithmica, 2001 30(4):527--562.
> 
> -Steve
>