Features based on quasi-sinudoidal modeling for tremor detection in Parkinsonian voice

This paper aims defining features to characterize Parkinsonian voice affected by tremor. It uses quasi-sinusoidal modelling of signals which assumes that speech signal is a sum of sinusoids with time-linearly varying instantaneous amplitudes and frequencies permits. The parameters of this model are calculated and their behavior is analyzed. The statistical analysis using box-plots permits to show the ability of this model to discriminate the Parkinsonian voice from the healthy voice.