Reissuable Biometrics through Image-Based Handwritten Signature Verification

The privacy invasion of the biometric technology is getting public concerns due to the fact that biometric characteristics are immutable. In other words, their compromise is permanent. Reissuable biometrics was devised to make the reissuable or replaceable of biometric templates possible once they are found compromised. Biometric Strengthening is a form of reissuable biometrics. It strengthens the biometric templates by transforming the original template values to form a new set of values through the Gaussian distribution. The performance of Biometric Strengthening is evaluated in three possible intrusion scenarios. Probabilistic neural network (PNN) is employed as classifier. The compatibility of Biometric Strengthening and PNN shows the potential of using them in real world application. The experiments are tested on own image-based handwritten signature data set due to the lack of benchmark database.