Parkinson’s Disease Progression Assessment From Speech

Abstract

ABSTRACT: In recent years the interest in the analysis of speech of Parkinson's disease patients have been increasing. Different approaches have been proposed in order to detect several voice problems in Parkinson's patients. However, a relative low number of the contributions are focused on tracking the state of the disease over the time. In this thesis is proposed a new methodology to assess the progression of the disease per speaker. The speech of the patients is modeled individually considering speech recordings captured in different sessions. Voice impairments are analyzed considering a particular model that reflects articulatory problems of the patients. The proposed approach shows that is possible to track the disease progression from speech with a relative high correlation value.