Title:Topology preserving thinning for cell complexes

Abstract:A skeleton is a synthetic representation of an object that retains many significant topological and geometric properties of the original object. The process of obtaining the skeleton is usually referred to as skeletonization or thinning. Thinning is a very active research area thanks to its central role in reducing the amount of information to be processed by pattern recognition algorithms.
Most thinning algorithms are claimed to be "topology preserving," even though a precise statement of what that means is usually left unaddressed. The first aim of this paper is to clarify this issue by using homology theory. Then, a simple topology preserving iterative thinning algorithm, based on acyclicity tables automatically produced with homology computations, is presented. Contrarily to the usual practice of using rule-based approaches, the generality of the proposed algorithm is proved theoretically. Moreover, the algorithm not only deals with cubical complexes, but it may be easily extended to process any other unstructured cell decomposition of the object, provided that cells have some fixed structure. The acyclicity tables for cubical and simplicial complexes and an open source implementation of the thinning algorithm that works with unstructured 3-dimensional simplicial complexes are provided to allow their immediate use in applications.