Hill-climbing attack based on the uphill simplex algorithm and its application to signature verification

Abstract

A general hill-climbing attack to biometric systems based on a modification of the downhill simplex algorithm is presented. The scores provided by the matcher are used in this approach to adapt iteratively an initial estimate of the attacked template to the specificities of the client being attacked. The proposed attack is evaluated on a competitive feature-based signature verification system over both the MCYT and the BiosecurID databases (comprising 330 and 400 users, respectively). The results show a very high efficiency of the hill-climbing algorithm, which successfully bypassed the system for over 90% of the attacks with a remarkably low number of scores needed. © 2011 Springer-Verlag.