Fifth International Workshop on Models and Analysis of Vocal Emissions for Biomedical Applications (MAVEBA 2007)

Florence, Italy
December 13-15, 2007

Detecting Pathology in the Glottal Spectral Signature of Female Voice

Pedro Gómez-Vilda, Roberto Fernández-Baíllo, R. Martínez, C. Muñoz, L. M. Mazaira, A. Álvarez, Juan Ignacio Godino-Llorente

Grupo de Informática Aplicada al Procesado de Señal e Imagen, Universidad Politécnica de Madrid, Boadilla del Monte, Madrid, Spain

Pathology detection by the analysis of voice remained a challenging objective during last years’ research. Former studies have shown that the glottal signature defined on the glottal source spectral density contains helpful hints in determining the healthy or pathological condition of the subject. Therefore a good reconstruction of the glottal source after removing the vocal tract is essential for the study. Nevertheless it may be shown that gender is also strongly influencing this glottal signature. Therefore comparisons and clustering between healthy and pathological glottal signatures should have into account the patient’s gender. Through the present paper a new scheme for pathology detection is presented based on a priori determining the subject’s gender. Results from a database of healthy and pathological subjects are given to contribute in this sense. A study case is presented as an example in visualizing the interest of the approach.
Index Terms. Glottal Signature, Gender Detection, Pathology Detection

Full Paper (reprinted with permission from Firenze University Press)

Bibliographic reference.  Gómez-Vilda, Pedro / Fernández-Baíllo, Roberto / Martínez, R. / Muñoz, C. / Mazaira, L. M. / Álvarez, A. / Godino-Llorente, Juan Ignacio (2007): "Detecting pathology in the glottal spectral signature of female voice", In MAVEBA-2007, 183-186.