Observatorio de I+D+i UPM

Descripción
In this contribution we study different speech modeling techniques to detect patients with severe OSA envisioning the future classification of patients according to their priority of need identifying the most severe cases and reducing medical costs. Hidden Markov Models (HMMs) are used for detecting voices of OSA patients. Specific acoustic properties of continuous speech are modeled attending to different linguistic contexts which reflect discriminative physiological characteristics found in OSA patients. Experimental results on over a database including both severe OSA and healthy speakers shows an 85% correct classification rate is achieved by using whole-sentence HMMs, outperforming previous schemes proposed in the literature.
Internacional	Si
DOI	10.1007%2F978-3-642-35292-8_13
Edición del Libro
Editorial del Libro	Springer
ISBN	978-3-642-35291-1
Serie
Título del Libro	Advances in Speech and Language Technologies for Iberian Languages
Desde página	121
Hasta página	128

Esta actividad pertenece a memorias de investigación

• Autor: Ana Montero Benavides (UPM)
• Autor: Rubén Fernández Pozo (UPM)
• Autor: Jose Luis Blanco Murillo (UPM)
• Autor: Luis Alfonso Hernandez Gomez (UPM)

• Creador: Grupo de Investigación: Grupo de Aplicaciones del Procesado de Señal (GAPS)
• Departamento: Señales, Sistemas y Radiocomunicaciones