Cancelable Biometrics: Securing Biometric Face Template

Biometric authentication system automatically identifies or verifies a person using physical, biological, and behavioral characteristics such as face, iris, fingerprints, hand geometry or voice. Compared to the traditional identification and verification methods (ID, card, password), biometrics is more convenient for users, can reduce fraud, and is more secure. In the past decade, biometric became an important standard of security, intelligence, and law enforcement. In a traditional system, if password is compromised user can easily change it. However, biometric for the users is limited and unique, and if user biometric is compromised, it is impossible or highly difficult to change it for a particular system. Cancellable biometric has recently emerged as solution of the problem, which allows replacing the compromised biometric template. The concept of cancelable biometric or cancelability is to transform a biometric data or feature into a new one so that users can change their single biometric template in same or another system. In this chapter, we present a biometric system where both signal and feature level transformations are used to achieve cancelability. The proposed cancelable system is established based on three steps. In the first step the two-fold random selections of the signals are made. Each fold is then randomly projected using random projection technique. In the second step, the Principal Components Analysis (PCA) is used to reduce the feature dimension of the randomly projected folds. After the PCA, a dimension reduction feature-based fusion using k-mean clustering is applied to create single template for face biometrics. In the third step, to enhance the discriminability, the Linear Discriminant Analysis is applied to the features. These features are then used in a classifier to get the final authentication performance.