A Cancelable Biometric Hashing for Secure Biometric Verification System

In this paper, we propose a secure and robust biometric hashing method. We use Radon transform, min-max quantizer and Reed-Muller decoder in this method. Our goal is to preserve privacy in a biometric recognition system while achieving the desirable accuracy rate. We do the verification using the hash values and hence increase the security of the system. In addition, we guarantee the irreversibility due to the properties of hashing methods. Moreover, if an attacker compromises the hash value, a new hash value can be reconstructed by changing the password. This ensures that the biometric hash values are cancelable. The simulation results demonstrate the efficiency of the proposed method. We achieve an equal error rate (EER) of 0.00145 on Multi Modal Verification for Teleservices and Security applications (M2VTS) face database. In addition, even in case the attacker compromises the secret key and the random number generator, we achieve an EER of 0.1185.