A Multimodal User Authentication Using Fuzzy Fusion of Face, Finger and Speech

Abstract

Use of biometrics in user authentication has increased tremendously in past few years due to its inherent advantages
over the conventional methods. Currently, most of the biometric system rely on single biometric identifier such as face, finger,
voice, retina, hand geometry, DNA, etc.. In general accuracy of these system ranges from 80-95% which is good for most of the
basic applications, but in case of high security surveillance system a higher degree of accuracy is desirable. To solve such
problem multi-modal biometrics systems are proposed where more than one biometrics are combined together for user
authentication. In this paper a multimodal biometric system is discussed where face, finger and voice are combined together
using fuzzy logic. The main advantages of the systems is that it overcomes the limitations of the single biometric identifier and it
is fast and accurate enough that it can be easily used in real time system. The obtained results clearly reveal that the
multimodal system provides very good accuracy in terms of user identification and authentication.