A secure Multimodal Biometric Authentication with Cryptographic key Management Using Double Random Phase Encoding

Multibiometric systems are more efficient and reliable than unibiometric systems as they can provide lower error rates as well as robustness against frauds and subsystem failures. However, the deployment of multibiometric systems in large-scale biometric applications increases the risk of users‟ privacy violation because once a multibiometric system is compromised; multiple biometric traits are disclosed to adversaries. As a result, protecting biometric templates stored in centralized databases of multibiometric systems has become a necessary prerequisite to allow wide-spread deployment of these systems. In this paper, we propose a biometric template protection method for securing image templates in multibiometric systems using the double random phase encoding (DRPE) scheme. DRPE is a well-known image encryption scheme and therefore it is more suited to secure image-based biometric templates. First, the proposed method encodes a randomly generated key as a binary image. Second, the phase components of two images captured from two different biometric modalities; namely, palmprint and fingerprint are convolved to produce a multi-biometric image of the same size as the binary image-encoded key. Finally, image-encoded key is encrypted using DRPE employing the multi-biometric image as a cipher key. During authentication, the encoded key is correctly recovered only if genuine biometric images are presented to the system; otherwise, the authentication process fails. Therefore, the proposed method can not only protect image-based biometric templates but also can provide a reliable means for securing cryptographic keys. Experimental results illustrate that the proposed method can secure both biometric templates and cryptographic keys without sacrificing the recognition accuracy of the underlying unprotected biometric recognition system.