Communication Dans Un Congrès Année : 2025

Diffusion-based spectral super-resolution of third octave acoustic sensor data: is privacy at risk ?

Résumé

Third octave spectral recording of acoustic sensor data is an effective way of measuring the environment. While there is strong evidence that slow (1s frame, 1 Hz rate) and fast (125ms frame, 8Hz rate) versions lead by-design to unintelligible speech if reconstructed, the advent of high quality reconstruction methods based on diffusion may pose a threat, as those approaches can embed a significant amount of a priori knowledge when learned over extensive speech datasets.

This paper aims to assess this risk at three levels of attacks with a growing level of a priori knowledge considered at the learning of the diffusion model, a) none, b) multi-speaker data excluding the target speaker and c) target speaker. Without any prior regarding the speech profile of the speaker (levels a and b), our results suggest a rather low risk as the worderror-rate both for humans and automatic recognition remains higher than 89%.

Fichier principal
Vignette du fichier
tailleurEusipco2025.pdf (733.55 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Licence

Dates et versions

hal-05096000 , version 1 (03-06-2025)
hal-05096000 , version 2 (21-07-2025)

Licence

Identifiants

  • HAL Id : hal-05096000 , version 2

Citer

Modan Tailleur, Chaymae Benaatia, Mathieu Lagrange, Pierre Aumond, Vincent Tourre. Diffusion-based spectral super-resolution of third octave acoustic sensor data: is privacy at risk ?. 33rd European Signal Processing Conference (EUSIPCO 2025), European Association for Signal Processing (EURASIP), Sep 2025, Palerme, Italy. pp.306. ⟨hal-05096000v2⟩
2482 Consultations
358 Téléchargements

Partager

  • More