Fuzzy Logic Weight Estimation in Biometric-Enabled Co-authentication Systems

Abstract : In this paper, we introduce a co-authentication system that combines password, biometric features (face, voice) in order to improve the false reject rate (FRR) and false accept rate (FAR) in Android smartphone authentication system. Since the system performance is often affected by external conditions and variabilities, we also propose a fuzzy logic weight estimation method which takes three inputs: password complexity, face image illuminance and audio signal-to-noise-ratio to automatically adjust the weights of each factor for the security improvement. The proposed method is evaluated using Yale [5] and Voxforge [1] Databases. The experimental results are very promising, the FAR is 0.4 % and FRR almost equal 0 % when the user remembers his password.