Cancelable Biometric System Based on the Quaternion Implementation of Double Random Phase Encoding
Mahmoud Nasr, Adam Piórkowski, Mohamed Essam, Fathi E. Abd El‐Samie · 2023
The majority of contemporary security solutions rely on biometrics. Unfortunately, these systems have been subjected to hacking attacks for a very long period. The biometrics contained in databases will be lost permanently if these databases have been breached and stolen. Therefore, the demand for innovative cancelable biometric solutions is great. The fundamental idea behind cancelable biometrics is to use a different biometric template constructed using a 1-way transform or an encryption technique to keep the original biometric templates secure and out of the system. In this research work, we depend on the basic idea of the Double Random Phase Encoding (DRPE) algorithm and improve its performance by using another auxiliary input biometric and apply the quaternion mathematics for implementation of the proposed technique. The suggested cancelable biometric recognition technique performs well in simulation experiments even when noise is present, particularly when applied for cancelable face recognition, where we get a high area under the Receiver Operating Characteristic (ROC) curve of 0.994 and a low Equal Error Rate (EER) of 0.00174.