Titre :

Speaker Identification Based On Vocal Cords Vibrations signal: Effect Of The Window

Auteur(s) :

Ishac, Dany [Auteur]
University of Balamand [Liban] [UOB]
Abche, Antoine [Auteur]
University of Balamand [Liban] [UOB]
Karam, Elie [Auteur]
University of Balamand [Liban] [UOB]
Nassar, Georges [Auteur]
INSA Institut National des Sciences Appliquées Hauts-de-France [INSA Hauts-De-France]
Matériaux et Acoustiques pour MIcro et NAno systèmes intégrés - IEMN [MAMINA - IEMN]
Debavelaere-Callens, Dorothée [Auteur]
INSA Institut National des Sciences Appliquées Hauts-de-France [INSA Hauts-De-France]
Matériaux et Acoustiques pour MIcro et NAno systèmes intégrés - IEMN [MAMINA - IEMN]

Titre de la revue :

International journal of digital information and wireless communications (IJDIWC)

Pagination :

58-62

Éditeur :

Hong Kong : Society of Digital Information and Wireless Communications,

Date de publication :

2018-03

ISSN :

2412-6551

Mot(s)-clé(s) en anglais :

"Vocal cords vibration frequency"
"STFT"
"Barlett"
"Blackman"
"Hamming"
"Hanning"
"Rectangular"
"collar"
"transducer"
"speaker identification"

Discipline(s) HAL :

Sciences de l'ingénieur [physics]
Sciences de l'ingénieur [physics]/Acoustique [physics.class-ph]
Sciences de l'ingénieur [physics]/Matériaux
Sciences de l'ingénieur [physics]/Micro et nanotechnologies/Microélectronique

Résumé en anglais : [en]

A new speaker identification technique was proposed and presented. It is based on the acquisition of the individual’s signal of its vocal cords’ vibrations by attaching a piezoelectric transducer ...
Lire la suite >A new speaker identification technique was proposed and presented. It is based on the acquisition of the individual’s signal of its vocal cords’ vibrations by attaching a piezoelectric transducer element to a collar and wrapping it around the neck. Having collected the signal, the Short Term Fourier Transform (STFT), in conjunction with a particular window, is applied on the signal to decompose it into its frequency components. Then, the resulting spectrogram is normalized, the noise is removed and the corresponding features are extracted for identification purposes. The performance of the developed approach in conjunction with various windows’ types (the Bartlett, the Blackman, the Hamming, the Hanning and the Rectangular windows) and window’s size is studied and evaluated quantitatively. The results show that the proposed approach using the Hamming window of size 64 yields the best accuracy in the identification of the desired individualsLire moins >

Langue :

Anglais

Vulgarisation :

Non

Collections :

• Institut d'Électronique, de Microélectronique et de Nanotechnologie (IEMN) - UMR 8520

Source :

Harvested from HAL