On 12/19/2010 5:10 AM, Warp wrote:
>    Sorry for asking, but for what possible reason would one want to
> convert a mesh into an isosurface? A conversion to the other direction
> would make sense and, in fact, be really useful (at least in terms of
> rendering speed). However, converting a mesh to an isosurface seems to
> be completely backwards. Why would one want to do that?

Why wouldn't one want to do it? If it could be done efficiently, a 
person could give erosion to a statue, or add rusty lumps to an object, 
at will.

The big question is how to do it? You could make an object pattern from 
the mesh, and convert that to an isosurface. But that always results in 
a lot of artifacts. Or you could average jittered copies of the object 
pattern together, but that renders very slowly and the artifacts are 
still present (though somewhat reduced). Or perhaps... you could start 
with a spherical function and push and pull space at each triangle thus 
distributing it into the mesh's shape. It's only a theory, and might 
only make sense for small meshes, but maybe it would work...

Sam

Post a reply to this message