[vtkusers] point cloud simplification question

[kevnull at phutureshock.com]

Apologizes for the cross post to those on the dev list, I mistakenly added
this as an addendum to a dev-related post.

I have what seems to be a simple question about thinning point clouds that
I have been struggling with lately, that is, extracting topologically
clean 2-surfaces from a CT scan that has some "thickness" to it. I would
like to use only the points on the "exterior of the sampled surface" to
eventually run marching cubes on (or whatever reconstruction
algorithm...), is there a filter or set of transformations that I can do
on the data to thin my data set to those points on the "outside surface"
of the point cloud?

Details images and the data set here:

http://www.phutureshock.com/mememap/poor-quality-meshes-and-vertex-normals-from-a-point-cloud-with-densitythickness/

My data is in the form of an .xyz file not images, but perhaps someone
knows of something that would work that is ITK-related if I transformed my
point cloud into a stack of images somehow? Just a thought.

Thanks in advance!
Best,
Kevin