Discriminant feature vectors for characterizing ailment cough vs. simulated cough

Cough is the powerful mechanism of human body to clear the central airways. Cough is often triggered by the mucus that drains down the back of the throat. An infection in the lungs or upper airway passages can cause cough. This paper describes the characteristics of ailment (cold) cough with respect to simulated (healthy) cough sound signals. Standard signal processing methods are used to extract the features from the acoustic signal of cough sounds that characterize the cough sounds. Analysis of cough sounds is carried out using the instantaneous fundamental frequency (F0) as a source feature, Mel frequency cepstral coefficients (MFCCs) as filter features, and signal energy as a combined feature of both the source of excitation and the vocal tract system. First four formants of the human speech production mechanism are also used for characterizing the ailment cough vs. simulated cough. Further a Support Vector Machine (SVM) classifier is used for the automated classification of the ailment cough sound from the simulated cough sound signal. Encouraging results are obtained. The proposed approach can have potential applications in assisted diagnosis of diseases, based upon cough sound signal analysis.