Matcher Performance-Based Score Level Fusion Schemes For Multi-modal Biometric Authentication System

Abstract

Multimodal systems improves the performance of the authentication system by fusing the physiological or behavioral characteristics of an individual. The fusion can be carried out in score or feature level fusion. This paper explains the multimodal biometric system against the unimodal system to overcome several demerits in the former system and to increase its recognition rate. It integrates ECG, face and fingerprint on score level fusion. Feature vectors were obtained after processing the signal as well as images from the databases FVC2002/2004, Face94 and Physionet (MIT-BIH Arrythmia) for extracting the features. Matching scores and individual accuracy were computed separately on each biometric trait. Since the matchers on these three biometric traits gave different values, Matcher performance based fusion scheme is suggested on the specified traits. The normalization of the scores is determined using OVEBAMM (Overlap extrema-based min-max) technique. The performance analysis of these traits in unimodal system and multimodal system is arrived and they were plotted with respect to the ROC (Receiver Operating Characteristic) curve. The overall accuracy rate was achieved up to 92.6%.