Score Level Fusion for Fingerprint, Iris and Face Biometrics

Abstract

Single biometric systems suffer from many challenges such as noisy data, non-universality and spoof attacks. Multimodal biometric systems can solve these limitations effectively by using two or more individual modalities. In this paper fusion of fingerprint, iris and face traits are used at score level in order to improve the accuracy of the system. Scores which obtained from the classifiers are normalized first using min-max normalization. Then sum, product and weighted sum rules are used to get fusion. Experimental results show that multimodal biometric systems outperform unimodal biometric systems and weighted sum rule gives the best results comparing with sum or product method.