Privacy preserving speech, face and fingerprint based biometrie authentication system using secure signal processing

Abstract

Biometrics represents the identity of individuals. Physical characteristics like voice, face, fingerprint, etc. are used to recognize individuals. Biometrics are used as a promising method for authentication, but use of these raw biometric data results in some privacy concerns. In this paper, we propose a system model for privacy preserving biometric authentication system for speech, face and fingerprint authentication. Signal Processing in Encrypted Domain (SPED) is a domain in signal processing that deals with the processing of encrypted data which preserve users' privacy and maintains security. The secure primitives, Secure Inner Product (SIP) and Secure Log Sum (SLS) are used for as authentication protocols. The performance of the system is evaluated using False Acceptance Rate (FAR) and False Rejection Rate (FRR) measures. The simulation results provided, shows good accuracy.