Biometric Verification by Fusing Hand Geometry and Palmprint

Abstract

This paper presents a biometric recognition system with fusion of hand geometry and palmprint of a human hand based on wavelet transform and statistical moments. The feature extraction module adopts the gradient direction (i.e., angle) and quadratic spline function of wavelet transform as the discriminating texture features in palmprint, and the statistical moments calculated from hand geometry. The system generates the palmprint feature codes using a binary gray encoding. The recognition rates up to 94.17%, 95.50%, 96.67%, and 98.33%, respectively, using different feature extraction methods may be achieved.