Talk:Topological skeleton

Merge with Medial Axis?
Shouldn't medial axis and topological skeleton be merged? Renato (talk) 16:56, 10 July 2008 (UTC)


 * There are many skeletonization methods. Medial axis is only one of them. We also have an article about straight skeleton, for instance. This article deserves to be separate, I think, but should concentrate more on the broad overview and comparison of methods suggested by a few of the external links. —David Eppstein (talk) 17:03, 10 July 2008 (UTC)


 * David, thanks for answering. But, medial axis is not a skeletonization method (medial axis transform is). The medial axis is the skeleton itself; I haven't seen different definitions for skeleton and medial axis; both can be characterized by "the centers of maximal balls" or "the centers of the bi-tangent balls", or "thin line that preserves the topology of the original shape". I think that medial axis and skeleton should be merged in a single page (unless there are different definitions for them in the literature), and then there can be different pages for different skeletonization methods.


 * By the way, the wikipage for topological skeleton (which is referenced in the definition of medial axis in its wikipage) is in a very bad shape; we have to do something about it. Not only is it tiny, it doesn't even give a definition of skeleton . A merger with the medial axis page, which is in much better shape, could be the first step, and then improve from there. What do you say? Renato (talk) 16:29, 11 July 2008 (UTC)


 * I think "thin line that preserves the topology of the original shape" is the right definition. "The centers of the bi-tangent balls" is not: it is too specific. In particular, the straight skeleton provides skeletonization methods that form thin lines preserving the topology of the original shape, but that do not pass through the centers of bi-tangent balls. That is why I don't think medial axis is the right merge target: it matches only the specific definition for that one skeletonization method, and does not cover skeletonization more broadly. —David Eppstein (talk) 17:09, 11 July 2008 (UTC)