An unrestricted attack based on diffusion models that can achieve both good transferability and imperceptibility.
-
Updated
Sep 4, 2024 - Python
An unrestricted attack based on diffusion models that can achieve both good transferability and imperceptibility.
Revisiting Transferable Adversarial Images (arXiv)
[TMLR 2023] as a featured article (spotlight 🌟 or top 0.01% of the accepted papers). In this study, we systematically examine the robustness of both traditional and learned perceptual similarity metrics to imperceptible adversarial perturbations.
Project Page (PDCL-Attack, ECCV 2024)
Project Page (FACL-Attack, AAAI 2024)
Add a description, image, and links to the transferable-attacks topic page so that developers can more easily learn about it.
To associate your repository with the transferable-attacks topic, visit your repo's landing page and select "manage topics."