Robust anti-fraud algorithms against malicious AI-generated face images and travel documents

• Research unit: INNOSUISSE
• Project number: 106.729 IP-ICT

Key words (English)

ICT, Deep learning

Abstract (English)

The project aim is twofold: i) develop robust defenses against deepfakes both for identity document images and selfie-videos, and ii) take advantage of the capabilities of generative AI to improve the state of the art in digital authentication by making algorithms more robust and balanced.