Skip to Main Content
Personal identity recognition is vital to information society. Multimodal biometric is proved to be a promising authentication technique. The fusion recognition algorithms were studied based on three biometric traits, hand vein, iris and fingerprint. The independent feature extraction and matching modules were designed for each unimodal biometric to output matching scores. The fusion operation was implemented at the matching score level after score normalization. Simple average fusion recognition algorithm was finally employed to realize the personal identity authentication. The experiments were carried out to verify availability of the proposed algorithm. The experiment results showed the multimodal biometric system had much higher recognition accuracy than any unimodal biometric system.
Date of Conference: 16-17 July 2011