Application of new alteration attack on biometric authentication systems

Abstract

Biometrics-based authentication systems are presented as an alternative solution to the traditional authentication techniques, these systems have more security advantages compared to a password as biometric data cannot be forgotten or wasted. However, the protection of biometric systems against different types of attacks is not yet guaranteed. While many studies are proposed in order to increase the security of biometric systems, several attacks are developed in order to exploit the weaknesses of these systems. In this paper, we present a new alteration attack on biometric authentication systems. We suppose that the impostor has altered image of the real user and he presents it as request in order to gain unlawful access to the system. Altered image can be recuperated from a fingerprint trace in case of fingerprint based systems or user photograph for attacking face based authentication systems. Next, we study the impact of alteration level on the matching score for fingerprint and face based biometric systems. Our analysis shows that both systems are vulnerable to the proposed attack and the alteration level has serious impact on the security of biometric systems. The impostor can have a matching score greater than 90% for the tested fingerprint based biometric system and 190/194 associations are matched for the tested face based biometric authentication system.