try these search terms:

"marching cubes"
"delauney triangulation"

also try http://gts.sourceforge.net

There are other methods :)

Quote:The convex hull algorithm sounds like it will give the appropriate points that are involved, but do you know how to go about computing the faces?

An appropriate convex hull algorithm will compute the faces; unfortunately, in 3d the implementation can be a little complicated.

There's a standard algorithm whose name escapes me at the moment. However, I've found that a somewhat simpler incremental algorithm works very well.

The first step is to initialize the convex hull with a more or less optimal simplex (tetrahedron in 3d):

1. Find two points some distance apart to form a segment
2. Find a point some distance from the segment to form a triangle
3. Find a point some distance from the triangle to form a tetrahedron

What 'some distance' means depends, but the idea is that the initial simplex be non-degenerate.

Then, for each point:

1. If it's inside the current hull, discard
2. Else find all faces whose support plane it is in front of
3. Discard these faces
4. Create a triangle between the new point and each exposed edge

It's been a while since I implemented this, and may not have described it correctly (I certainly left out some details). There's also the issue of robustness - there are ways to deal with this problem if you need exact results.