In What Way Is It Possible to Impersonate You Bypassing Fingerprint Sensors?

In this study, we want to determine if an attacker with different a priori information on the sensor parameters during the enrollment process can impersonate an individual. We have tested different apriorism such as the fingerprint class, sensor type, image resolution and the number of minutiae. Two different matching algorithms have been used on specific designed databases generated with SFinge. Two attack scenarii have been tested. Obtained results show that the knowledge of fingerprint class and resolution influence the attack success.