Sixth International Workshop on Models and Analysis of Vocal Emissions for Biomedical Applications (MAVEBA 2009)
This paper focuses on the automatic detection of speech pathologies by exploiting the estimation of the glottal source. Three methods of estimation are compared and time and spectral features are extracted. The relevancy of these features is assessed by means of information theory-based measures. This allows an intuitive interpretation in terms of discrimination power and redundancy between the features. It is discussed which features are informative or complementary for detecting voice pathologies and the glottal source estimation methods are compared.
Index Terms. Voice Pathology, Glottal Source, Mutual Information
Full Paper (reprinted with permission from Firenze University Press)
Bibliographic reference. Dubuisson, T. / Drugman, T. / Dutoit, Thierry (2009): "On the mutual information of glottal source estimation techniques for the automatic detection of speech pathologies", In MAVEBA-2009, 53-56.