Palmprint recognition combining LBP and cellular automata

Abstract

In this paper, we present a palmprint recognition method which combines local binary pattern (LBP) and cellular automata. The LBP descriptor is proposed as a unifying texture model that describes the formation of a texture with micro-textons and their statistical placement rules. Because texture is one of the most importent features in palmprint image, so we think the features based on LBP will be good discriminative for palmprint identification. Cellular automata can be generally described as discrete dynamic systems completely defined by a set of rules in a local neighborhood. In this paper, we use cellular automata to extract features as the part of feature vector. The experiments conducted on Polytechnic University Palmprint Database I demonstrates the effectiveness of proposed method. © 2010 Springer-Verlag Berlin Heidelberg.