Evaluated the robustness of a deep face recognition model (InceptionResNetV1) against adversarial attacks. Tested multiple attack types, analyzed transferability, and implemented a defense system using specialized detectors to improve security while preserving accuracy.
detector cybersecurity adversarial-machine-learning adversarial-attacks vggface2 deep-face-recognition adversarial-detection carlini-wagner-attack adversarial-robustness-toolbox fgsm-attack pgd-attack bim-attack security-evaluation-curve
-
Updated
Jan 16, 2026 - Python