Geometric Invariant Shape Classification Using Hidden Markov Model

Abstract

In this paper we propose a novel approach for geometric shape classification by using shape simplification and discrete Hidden Markov Model (HMM). The HMM is constructed using the landmark points obtained from the shape simplification for each shape image in the dataset. Some useful strategies have been employed for the constructed HMM for geometric shape classification. Experimental results based on the common MPEG7 CE shapes database shows that our proposed method can achieve very good accuracy in different kinds of shapes.