Fusion of Gaussian mixture densities for face and ear biometrics using support vector machines

Abstract

This paper presents a multimodal biometric system for face and ear biometrics which convolves face and ear images with Gabor wavelet filters for extracting enhanced Gabor features from the corresponding images which are characterized by spatial frequency, spatial locality and orientation. Gaussian Mixture Model (GMM) is applied to the Gabor responses for measurements and Expectation Maximization algorithm is used to estimate density parameters in GMM. It produces two sets of feature sets which are fused using Support Vector Machines. Experiments on two different databases reveal its usefulness towards robust multimodal fusion. © 2010 Springer-Verlag Berlin Heidelberg.