Log-normal Mutations and their Use in Detecting Surreptitious Fake Images - Archive ouverte HAL
Pré-Publication, Document De Travail Année : 2024

Log-normal Mutations and their Use in Detecting Surreptitious Fake Images

Résumé

In many cases, adversarial attacks are based on specialized algorithms specifically dedicated to attacking automatic image classifiers. These algorithms perform well, thanks to an excellent ad hoc distribution of initial attacks. However, these attacks are easily detected due to their specific initial distribution. We therefore consider other black-box attacks, inspired from generic black-box optimization tools, and in particular the log-normal algorithm. We apply the log-normal method to the attack of fake detectors, and get successful attacks: importantly, these attacks are not detected by detectors specialized on classical adversarial attacks. Then, combining these attacks and deep detection, we create improved fake detectors.
Fichier principal
Vignette du fichier
main.pdf (10.29 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04710415 , version 1 (26-09-2024)

Identifiants

  • HAL Id : hal-04710415 , version 1

Citer

Ismail Labiad, Thomas Bäck, Pierre Fernandez, Laurent Najman, Mariia Zameshina, et al.. Log-normal Mutations and their Use in Detecting Surreptitious Fake Images. 2024. ⟨hal-04710415⟩
24 Consultations
6 Téléchargements

Partager

More