An integrated top-down/bottom-up approach to speaker diarization

Abstract : Most speaker diarization systems fit into one of two cat- egories: bottom-up or top-down. Bottom-up systems are the most popular but can sometimes suffer from instability from merging and stopping criteria difficulties. Top-down systems deliver competitive results but are particularly prone to poor model initialization which often leads to large variations in performance. This paper presents a new integrated bottom-up/top-down approach to speaker diarization which aims to harness the strengths of each system and thus to improve performance and stability. In contrast to previous work, here the two systems are fused at the heart of the segmentation and clustering stage. Experimental results show improvements in speaker diarization performance for both meeting and TV-show domain data indicating increased intra and inter-domain stability. On the TV-show data in particular, an average relative improvement of 32% DER is obtained.
Type de document :
Communication dans un congrès
Interspeech 2010, September 26-30, Makuhari, Japan, Sep 2010, Makuhari, Tokyo, Japan. Interspeech 2010, September 26-30, Makuhari, Japan, 2010
Liste complète des métadonnées

Littérature citée [14 références]  Voir  Masquer  Télécharger

https://hal.archives-ouvertes.fr/hal-00601390
Contributeur : Simon Bozonnet <>
Soumis le : vendredi 17 juin 2011 - 15:32:37
Dernière modification le : mardi 18 septembre 2018 - 17:18:03
Document(s) archivé(s) le : dimanche 18 septembre 2011 - 02:28:11

Fichier

interspeech2010_1_.pdf
Fichiers éditeurs autorisés sur une archive ouverte

Identifiants

  • HAL Id : hal-00601390, version 1

Collections

Citation

Simon Bozonnet, Nicholas Evans, Corinne Fredouille, Dong Wang, Raphaël Troncy. An integrated top-down/bottom-up approach to speaker diarization. Interspeech 2010, September 26-30, Makuhari, Japan, Sep 2010, Makuhari, Tokyo, Japan. Interspeech 2010, September 26-30, Makuhari, Japan, 2010. 〈hal-00601390〉

Partager

Métriques

Consultations de la notice

176

Téléchargements de fichiers

120