Thesis on Automatic Speech Recognition
"Analyzing the Impact of Acoustic Features on Parkinson's Disease Severity Prediction"This project explores how different speech features, such as Mel-Frequency Cepstral Coefficients (MFCC), Chroma, and Delta features, can be used to predict Parkinson's disease (PD) severity. A Transformer-based model is applied to learn patterns in Italian speech recordings, with particular focus on temporal changes captured by Delta features. The goal is to identify which features are most informative for predicting disease severity. Insights from this work could support the development of reliable diagnostic tools to improve early detection, track disease progression, and enhance treatment strategies for PD patients.
GitHub Repository Link