A hitchhiker's guide to white-box neural network watermarking robustness - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

A hitchhiker's guide to white-box neural network watermarking robustness

Résumé

The present study deals with white-box Neural Network (NN) watermarking and focuses on the robustness property. The first contribution consists of formalizing neuron permutation as a geometric attack, thus demonstrating the very existence of this class of attacks for NN watermarking. The second contribution consists in devising and demonstrating the effectiveness of the corresponding counterattack. As a side result, the possibility of extending NN white-box watermarking scope beyond image classification is brought to light. The experimental study considers three state-of-the-art methods, four NN models, three tasks (image classification, segmentation, and video coding), and five types of attacks. We underline that none of the existing methods is robust against the geometric attack, and using the counterattack advanced in this paper effectively ensures the robustness.
Fichier principal
Vignette du fichier
EUVIP23_permutation_watermark.pdf (800.78 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04230306 , version 1 (05-10-2023)

Identifiants

  • HAL Id : hal-04230306 , version 1

Citer

Carl de Sousa Trias, Mihai P Mitrea, Enzo Tartaglione, Attilio Fiandrotti, Marco Cagnazzo, et al.. A hitchhiker's guide to white-box neural network watermarking robustness. The 11th European Workshop on Visual Information Processing (EUVIP ), Sep 2023, Gjovik, Norway. ⟨hal-04230306⟩
35 Consultations
34 Téléchargements

Partager

Gmail Facebook X LinkedIn More