Optimizing Secure Selective Face Template Generation Exchange over Open Network

Abstract Many modern authentication systems utilize human biometrics instead of traditional passwords and security codes to overcome their limitations. To enhance security, cancellable biometric transformations are employed to make it more difficult to retrieve the original biometric data. This paper introduces a new method for cancellable face recognition using Elliptic Curve Cryptography (ECC) along with selective biometric techniques. The goal of this approach is to strengthen the security of biometric traits, protect against potential attacks, and facilitate the safe transfer of biometric templates over open networks and within low-storage databases. By generating cancellable biometric patterns from the originals, this system can effectively manage access while addressing common issues associated with biometric systems, such as sensor data corruption, missing characters, under-representation, overshoot, and incompleteness. Additionally, a multi-modal biometric identification system can lower Failure-To-Capture (FTC) and Failure-To-Enroll (FTE) rates, providing strong protection against counterfeiting. To improve the implementation of multi-modal biometric traits, the time-consuming ECC multiplication operation is circumvented by using selective or partial biometrics as the template for each user, making it more suitable for real-time applications. Importantly, the proposed framework ensures complete distortion and encryption of unique biometric traits to protect them from unauthorized access. The efficiency and robustness of this approach have been validated using two sets of face biometric databases. Performance was evaluated using Receiver Operating Characteristic (ROC) curves and correlation scores, with simulation results demonstrating the effectiveness and potential of the proposed method.