[vtkusers] Fast implicit surface from curves

Elie Jamaa elie at dneg.com
Thu Oct 12 10:29:57 EDT 2000


Thanks very much, that sounds very encouraging. I'll get on it then. My main
worry though is the tesselation bit because I've got the feeling that the
general approach is to "grid" the isosurface from the X,Y,Z planes but
wouldn't it be more optimized to have a triangulation locally aligned to the
curve's tangent or something of the kind ? The point would be to have a
higher triangulation in the highest curvature areas.

Thanks again,  ;)


> The algorithm could be made probably 1-2 orders of magnitude
> faster by coding it specifically for curves. If you'd like to take a
> crack at it, basically what you have to do is compute a distance
> value surrounding the curve in a volume, then isosurface the
> distance value.

More information about the vtkusers mailing list