Accéder directement au contenu Accéder directement à la navigation
Communication dans un congrès

OVERLAP-AWARE LOW-LATENCY ONLINE SPEAKER DIARIZATION BASED ON END-TO-END LOCAL SEGMENTATION

Juan Manuel Coria 1 Hervé Bredin Sahar Ghannay Sophie Rosset
1 ILES - Information, Langue Ecrite et Signée
LIMSI - Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur
Abstract : We propose to address online speaker diarization as a combination of incremental clustering and local diarization applied to a rolling buffer updated every 500ms. Every single step of the proposed pipeline is designed to take full advantage of the strong ability of a recently proposed end-to-end overlapaware segmentation to detect and separate overlapping speakers. In particular, we propose a modified version of the statistics pooling layer (initially introduced in the x-vector architecture) to give less weight to frames where the segmentation model predicts simultaneous speakers. Furthermore, we derive cannot-link constraints from the initial segmentation step to prevent two local speakers from being wrongfully merged during the incremental clustering step. Finally, we show how the latency of the proposed approach can be adjusted between 500ms and 5s to match the requirements of a particular use case, and we provide a systematic analysis of the influence of latency on the overall performance (on AMI, DIHARD and VoxConverse).
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-03375330
Contributeur : Juan Manuel Coria Connectez-vous pour contacter le contributeur
Soumis le : mardi 12 octobre 2021 - 16:38:17
Dernière modification le : mardi 16 novembre 2021 - 04:01:26

Fichier

this-paper.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-03375330, version 1

Citation

Juan Manuel Coria, Hervé Bredin, Sahar Ghannay, Sophie Rosset. OVERLAP-AWARE LOW-LATENCY ONLINE SPEAKER DIARIZATION BASED ON END-TO-END LOCAL SEGMENTATION. IEEE Automatic Speech Recognition and Unserstanding Workshop, Dec 2021, Cartagena, Colombia. ⟨hal-03375330⟩

Partager

Métriques

Consultations de la notice

57

Téléchargements de fichiers

25