Kévin Carta, André Huynh, Stéfane Mouille, Nadia El Mrabet, Claude Barral, Sébastien Brangoulo

Proceedings of the 14th International Multi-Conference on Complexity, Informatics and Cybernetics: IMCIC 2023, pp. 105-112 (2023); https://doi.org/10.54808/IMCIC2023.01.105

Kévin Carta
CLR Labs, La Ciotat, France

André Huynh
CLR Labs, La Ciotat, France

Stéfane Mouille
CLR Labs, La Ciotat, France

Nadia El Mrabet
Ecole des Mines de St Etienne, Gardanne, France

Claude Barral
Bactech, Aix-en-Provence, France

Sébastien Brangoulo
Unissey, France, France

Cite this paper as:
Carta, K., Huynh, A., Mouille, S., El Mrabet, N., Barral, C., Brangoulo, S. (2023). How Video Injection Attacks Can Even Challenge State-of-the-Art Face Presentation Attack Detection Systems. In N. Callaos, S. Hashimoto, N. Lace, B. Sánchez, M. Savoie (Eds.), Proceedings of the 14th International Multi-Conference on Complexity, Informatics and Cybernetics: IMCIC 2023, pp. 105-112. International Institute of Informatics and Cybernetics. https://doi.org/10.54808/IMCIC2023.01.105

Owing to the digitization of our society, recently rushed by COVID 19 pandemic, remote biometrics, particularly facial recognition, are emerging in plenty of applications that require authentication at distance. These sensitive applications, mostly used in banking, governmental, or insurance domains, are threatened by the well-known presentation attack threat but also by video injection attacks (modification of the images taken by the camera by the attacker), which are now more accessible to fraudsters as devices are not scrutinized and under their control. We will see in this paper that even a simple photo can fool state-of- the-art biometric attack detection systems. Thus, we will give recommendations to avoid the usage of video injection attacks by malicious applicants.