A Multimodal Hand Vein Biometric based on Score Level Fusion

Abstract

Traditional security methods have largely been overtaken by biometrics. Vein pattern characteristics have become the forefront of biometric research because of its uniqueness, stability and immunity to frauds. Researchers are pioneering methods of processing and matching vein patterns. For the last few years, hand vein unimodal biometric has been explored. However, to address the challenges such as intra-class variations, unacceptable error rates and noisy data posed by the latter, multibiometrics has to be developed. Motivated by the fact that multibiometrics improves the accuracy of biometric system, a hand vein biometric comprising of dorsal and palmar vein has been implemented in this work. First, individual scores are generated by the individual matchers and are used for testing the biometric system. These scores are then fused using score level fusion, which is easy to access and combine the scores obtained from the different modalities. Unimodal biometric performance has been compared with hand multimodal biometrics.