Spoofing Speaker Verification With Voice Style Transfer And Reconstruction Loss - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2021

Spoofing Speaker Verification With Voice Style Transfer And Reconstruction Loss

Frauder la vérification de locuteur avec du transfert de style de voix et une loss de reconstruction

Résumé

In this paper we investigate a template reconstruction attack against a speaker verification system. A stolen speaker embedding is processed with a zero-shot voice-style transfer system to reconstruct a Mel-spectrogram containing as much speaker information as possible. We assume the attacker has a black box access to a state-of-the-art automatic speaker verification system. We modify the AutoVC voice-style transfer system to spoof the automatic speaker verification system. We find that integrating a new loss targeting embedding reconstruction and optimizing training hyper-parameters significantly improves spoofing. Results obtained for speaker verification are similar to other biometrics, such as handwritten digits or face verification. We show on standard corpora (VoxCeleb and VCTK) that the reconstructed Mel-spectrograms contain enough speaker characteristics to spoof the original authentication system.
Fichier principal
Vignette du fichier
Spoofing_Speaker_Verification_With_Voice_Style_Transfer_And_Reconstruction_Loss.pdf (2.2 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03356005 , version 1 (27-09-2021)

Identifiants

  • HAL Id : hal-03356005 , version 1

Citer

Thomas Thebaud, Gaël Le Lan, Anthony Larcher. Spoofing Speaker Verification With Voice Style Transfer And Reconstruction Loss. Workshop on Information Forensics and Security, Dec 2021, Montpellier, France. ⟨hal-03356005⟩
134 Consultations
162 Téléchargements

Partager

Gmail Facebook X LinkedIn More