Image-based Multimodal Biometric Authentication Using Double Random Phase Encoding

This paper presents an image-based multimodal biometric authentication scheme that utilizes a popular image encryption technique, known as double random phase encoding (DRPE). The proposed scheme aims at protecting biometric templates against unauthorized disclosure without deteriorating the recognition accuracy. Unlike conventional feature-based biometric authentication methods, the proposed scheme uses phase-based image matching of images captured from the employed biometric modalities; namely, palmprint and fingerprint, in order to enhance both accuracy and security of the suggested recognition system. A palmprint image is used as a secret key to encrypt a fingerprint image, captured from the same user, using DRPE. The encrypted fingerprint image can be successfully decrypted only if the phaseonly-correlation between enrollment and authentication palmprint images is sufficiently high (i.e. the two images belonging to the same user). Then, the decrypted image is matched against a fresh fingerprint image, also captured during verification, and the overall authentication process succeeds if the matching result exceeds a predefined threshold. The experimental results illustrate the efficacy of the proposed scheme and show that it can improve the security of biometric data without deteriorating the recognition accuracy.