91桃色

Abstract

The larynx cancer is the most common pathology among the smokers. Otolaryngologists use the treatment of this pathology that is usually the extirpation of the larynx and so its vocal folds, so the patients produce their voice directly from the esophagus. The objective parameters of characterization for these kind of esophageal voices (as pitch, jitter, shimmer or HNR) get out of the ranges normality due to the low quality of the sound itself and its intelligibility. In this talsk is introduced a new kind of technological algorithm which corrects the differences between normal and oesophageal parameters by different techniques as Wavelet Transform, Kalman filtering, LPC speech model and give some ideas about how to improve the quality by speech signal processing.

Biography

Bego?a Garc?a Zapirain was born in San Sebastian (Spain) in 1970. She graduated in Telecommunications Engineering, specializing in Telematics, at the Basque Country University in 1994. In 2003 she defended her doctoral thesis in the pathological speech digital processing field. After many years working for ZIV Company, in 1997 she joined the University of Deusto faculty as a lecturer in signal theory and electronics. She has led the Telecommunications Department at University of Deusto since 2002. In 2001, along with Javier Vicente S?ez, she created the Advanced Signal Processing (PAS) research group at the same university, playing the role of head researcher. Member of IEEE and EURASIP, she is nowadays part of the ISIVC and ISSPIT organization committee.