AVA: Inconspicuous Attribute Variation-based Adversarial Attack bypassing DeepFake Detection. (arXiv:2312.08675v1 [cs.CV])

While DeepFake applications are becoming popular in recent years, their
abuses pose a serious privacy threat. Unfortunately, most related detection
algorithms to mitigate the abuse issues are inherently vulnerable to
adversarial attacks because they are built atop DNN-based classification
models, and the literature has demonstrated that they could be bypassed by
introducing pixel-level perturbations. Though corresponding mitigation has been
proposed, we have identified a new attribute-variation-based adversarial attack
(AVA) that perturbs the latent space via a combination of Gaussian prior …

abuse adversarial adversarial attacks algorithms applications attack attacks bypassing classification deepfake deepfake detection detection literature popular privacy serious threat vulnerable