Baselines and Protocols for Household Speaker Recognition - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Communication Dans Un Congrès Année : 2022

Baselines and Protocols for Household Speaker Recognition

Résumé

Speaker recognition on household devices, such as smart speakers, features several challenges: (i) robustness across a vast number of heterogeneous domains (households), (ii) short utterances, (iii) possibly absent speaker labels of the enrollment data (passive enrollment), and (iv) presence of unknown persons (guests). While many commercial products exist, there is less published research and no publicly-available evaluation protocols or open-source baselines. Our work serves to bridge this gap by providing an accessible evaluation benchmark derived from public resources (VoxCeleb and ASVspoof 2019 data) along with a preliminary pool of open-source baselines. This includes four algorithms for active enrollment (speaker labels available) and one algorithm for passive enrollment.
Fichier principal
Vignette du fichier
Manuscript_Household.pdf (388.7 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03846180 , version 1 (10-11-2022)

Identifiants

Citer

Alexey Sholokhov, Xuechen Liu, Md Sahidullah, Tomi Kinnunen. Baselines and Protocols for Household Speaker Recognition. The Speaker and Language Recognition Workshop (Odyssey 2022), Jun 2022, Beijing, China. pp.185-192, ⟨10.21437/Odyssey.2022-26⟩. ⟨hal-03846180⟩
11 Consultations
147 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More