An m-health system for the estimation of voice disorders

In recent years, the prevalence of voice disorders has been increasing dramatically, mainly due to unhealthy lifestyles and voice abuse. An impairment in the ability to produce the sound of the voice is indicated by the medical term Dysphonia. This disorder impacts on the quality of life of the population in general, but especially of professional voice users, like teachers or singers, for whom dysphonia is more common. In this paper we present an accurate and robust methodology for the estimation of the Fundamental Frequency (F₀) developed in a mobile application and able to perform a simple and fast voice screening. The mobile app acquires and analyzes vocal signals requiring of the user only a vocalization of the vowel /a/. Automatically, the app indicates to the user his/her own voice status evaluating the value of F₀, and discriminating a pathological voice from healthy one. To estimate the classification accuracy of the implemented methodology, we have performed several experimental tests on an available database and we have compared the results with other algorithms in the literature.