Cancelable Speaker Identification System Based on Optical-Like Encryption Algorithms br

Biometric authentication is a rapidly growing trend that is gainingincreasing attention in the last decades. It achieves safe access to systems usingbiometrics instead of the traditional passwords. The utilization of a biometric inits original format makes it usable only once. Therefore, a cancelable biometrictemplate should be used, so that it can be replaced when it is attacked. Cancelablebiometrics aims to enhance the security and privacy of biometric authentication.Digital encryption is an efficient technique to be used in order to generate cancel-able biometric templates. In this paper, a highly-secure encryption algorithm isproposed to ensure secure biometric data in verification systems. The consideredbiometric in this paper is the speech signal. The speech signal is transformed intoits spectrogram. Then, the spectrogram is encrypted using two cascaded opticalencryption algorithms. Thefirst algorithm is the Optical Scanning Holography(OSH) for its efficiency as an encryption tool. The OSH encrypted spectrogramis encrypted using Double Random Phase Encoding (DRPE) by implementingtwo Random Phase Masks (RPMs). After the two cascaded optical encryptionalgorithms, the cancelable template is obtained. The verification is implementedthrough correlation estimation between enrolled and test templates in theirencrypted format. If the correlation value is larger than a threshold value, the useris authorized. The threshold value can be determined from the genuine and impos-ter correlation distribution curves as the midpoint between the two curves. Theimplementation of optical encryption is adopted using its software rather thanthe optical setup. The efficiency of the proposed cancelable biometric algorithmis illustrated by the simulation results. It can improve the biometric data securitywithout deteriorating the recognition accuracy. Simulation results give close-to-zero values for the Equal Error Rate (EER) and close-to-one values for the Area underReceiver Operator Characteristic (AROC) curve