Digital-to-Physical Transfer of Adversarial Patches for Aerial Vehicle Detection
Researchers have developed a method to create physical adversarial patches that can fool deep neural network-based aerial vehicle detectors. These patches are optimized digitally with constraints for printability and smoothness, then printed and tested in real-world conditions. While digital optimization showed high effectiveness for one patch configuration, another proved more robust in physical environments, highlighting practical security vulnerabilities in aerial detection systems. AI
IMPACT Highlights practical vulnerabilities in AI-powered aerial surveillance and detection systems, necessitating improved robustness measures.
