Détection automatique des instants de fermeture glottale dans les voix pathologiques

Résumé : Speech processing is an important thematic of the sciences of the engineer allying signal processing and medical knowledge. Speech is used as an information vector for many industrial applications, such as speech recognition, man-machine interfaces and many other applications. The study of the speech could allow the diagnosis of a disease. Many diseases has voice disorders as symptoms. Some of them are due to a malfunction of the vocal cords. Speech-based methods have been developed in recent years to identify instants when vocal cords come into contact, these instants are called GCIs (Glottal Closure Instant). The ground truth is obtained via electroglottography (EGG). The EGG provides an image signal of the movement of the vocal cords. From this signal several methods allow us to extract the GCIs automatically. These methods give different and sometimes false results for non-pathological voices and is probably not suitable for pathological voices. It is therefore difficult to obtain a reliable ground truth. The ground truth is nevertheless essential to be able to develop and study methods of detection of GCIs. This is exactly the main subject I had to work on during my internship.
Type de document :
Mémoires d'étudiants -- Hal-inria+
Traitement du signal et de l'image. 2017
Liste complète des métadonnées

Littérature citée [8 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-01627875
Contributeur : Khalid Daoudi <>
Soumis le : jeudi 2 novembre 2017 - 15:13:16
Dernière modification le : vendredi 3 novembre 2017 - 01:11:09
Document(s) archivé(s) le : samedi 3 février 2018 - 13:53:12

Fichier

Rapport-Robin.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01627875, version 1

Collections

Citation

Quentin Robin. Détection automatique des instants de fermeture glottale dans les voix pathologiques. Traitement du signal et de l'image. 2017. 〈hal-01627875〉

Partager

Métriques

Consultations de la notice

97

Téléchargements de fichiers

36