Observatorio de I+D+i UPM

Descripción
Gender detection from running speech is a very important objective to improve efficiency in tasks as speech or speaker recognition, among others. Traditionally gender detection has been focused on fundamental frequency (f0) and cepstral features derived from voiced segments of speech. The methodology presented here discards f0 as a valid feature because its estimation is complicate, or even impossible in unvoiced fragments, and its relevance in emotional speech or in strongly prosodic speech is not reliable. The approach followed consists in obtaining uncorrelated glottal and vocal tract components which are parameterized as mel-frequency coefficients. K-fold and cross-validation using QDA and GMM classifiers showed detection rates as large as 99.77 in a gender-balanced database of running speech from 340 speakers.
Internacional	Si
JCR del ISI	No
Título de la revista	Lecture Notes on Artificial Intelligence
ISSN	0302-9743
Factor de impacto JCR
Información de impacto
Volumen	7911
DOI
Número de revista
Desde la página	25
Hasta la página	32
Mes	SIN MES
Ranking

Esta actividad pertenece a memorias de investigación

• Autor: Cristina Elena Muñoz Mulas (UPM)
• Autor: Rafael Martinez Olalla (UPM)
• Autor: Pedro Gomez Vilda (UPM)
• Autor: Agustin Alvarez Marquina (UPM)
• Autor: Luis Miguel Mazaira Fernandez (UPM)

• Creador: Grupo de Investigación: Informática Aplicada al Procesado de Señal e Imagen
• Centro o Instituto I+D+i: Centro de tecnología Biomédica CTB
• Departamento: Arquitectura y Tecnología de Sistemas Informáticos