Object recognition based on skeleton

Abstract

In computer vision, representing graphics with important information can reduce the workload of recognition. As an important research content of computer graphics representation, skeleton features are simple in structure, so images can be matched effectively and quickly. Skeleton provides an effective and compact representation of an image object by reducing its dimensionality to a skeleton while preserving the topologic and geometric properties of the object, so using the skeleton to identify the shape has certain advantages. In this paper, we present the key processes of the object recognition based on the skeleton.