AN OPEN-SOURCE SPEAKER GENDER DETECTION FRAMEWORK FOR MONITORING GENDER EQUALITY

Abstract : This paper presents an approach based on acoustic analysis to describe gender equality in French audiovisual streams, through the estimation of male and female speaking time. Gender detection systems based on Gaussian Mixture Models , i-vectors and Convolutional Neural Networks (CNN) were trained using an internal database of 2,284 French speakers and evaluated using REPERE challenge corpus. The CNN system obtained the best performance with a frame-level gender detection F-measure of 96.52 and a hourly gender speaking time percentage error bellow 0.6%. It was considered reliable enough to realize large-scale gender equality descriptions. The proposed gender detection system has been packaged as an open-source framework.
Type de document :
Communication dans un congrès
IEEE International Conference on Acoustic Speech and Signal Processing, Apr 2018, Calgary, Canada
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01927560
Contributeur : Anthony Larcher <>
Soumis le : lundi 19 novembre 2018 - 22:57:53
Dernière modification le : vendredi 23 novembre 2018 - 01:12:42

Fichier

ddoukhan_icassp_2018.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01927560, version 1

Citation

David Doukhan, Jean Carrive, Félicien Vallet, Anthony Larcher, Sylvain Meignier. AN OPEN-SOURCE SPEAKER GENDER DETECTION FRAMEWORK FOR MONITORING GENDER EQUALITY. IEEE International Conference on Acoustic Speech and Signal Processing, Apr 2018, Calgary, Canada. 〈hal-01927560〉

Partager

Métriques

Consultations de la notice

18

Téléchargements de fichiers

20