Towards realistic fingerprint presentation attacks: The ScreenSpoof method

Casula, Roberto;Micheletto, Marco;Orru', Giulia;Marcialis, Gian Luca;Roli, Fabio

2022-01-01

---

Abstract

Using mobile devices, people leave latent fingerprints on the screen that can be photographed and used as a mold to fabricate high-quality spoofs. In this paper, we analysed the threat level of realistic attacks using snapshot pictures of latent fingerprints against presentation attack detection systems. Our evaluation compares a semi-consensual acquisition, given by a voluntary pressure of the finger on the screen, with a completely non-consensual one, given by the normal use of the device.