Parkinson’s Speech Detection Using YAMNet

Abstract

Parkinson's disease (PD) is a neurological condition that progressively impairs speech, balance, and mobility. For PD to be effectively managed and treated, early detection and diagnosis are essential. As a non-invasive method for early disease detection, speech signals can be used. Manual speech signal analysis, however, can be laborious and error prone. Thus, methods based on Deep Learning (DL) have been suggested to streamline the procedure and increase precision. YAMNet, a deep-learning model for audio categorization which is computationally efficient, was used to extract features from the PD dataset. The effectiveness and accuracy of the predictions were evaluated. Using the analysis of speech signals, the study sought to create a precise and effective tool for the early identification and management of PD. In this study, Parkinson detection is achieved using YAMNet model. The accuracy of around 82% was encouraging, highlighting the potential of employing speech signals as a PD diagnostic tool.