An empirical study of multi-mode biometric systems using face and fingerprint

Abstract

Recent deployments of biometrics application systems have revealed various limitations of the person authentication technology based on a single biometric feature. This paper empirically studies a multi-mode system by integrating face and fingerprint recognition to overcome the limitations of individual single-mode biometric systems and to achieve an increase in performance. This also proposes a framework of the performance evaluation of multiple biometric systems. The multi-mode system has been tested and evaluated by fusion of confidence levels in scenario and operational tests. The results of this research on combining two biometric modalities demonstrate a considerable amount of improvement in EER (Equal Error Rate) and ROC (Receiver Operating Characteristic) curves than single-mode biometric systems. © Springer-Verlag 2004.