Finding and ranking basic structures on complex line patterns

Abstract

The skeleton of a digital shape is a line pattern consisting of two basic structures, the branches and the loops, which are interleaved through the branch points. An iterative parallel algorithm is illustrated to find these basic structures, no matter how complex the skeleton is. In this way, the digital shape can be decomposed into regions, which reflect the topological properties of the corresponding skeleton structures. A key point of the proposed method is the notion of interiority degree, which allows one to order skeleton structures in a hierarchical way