Acoustic Voice Analysis for Parkinson’s Detection
Developing a non-invasive acoustic screening framework designed to identify early-stage neurodegenerative speech degradation through vocal signal analysis. By extracting micro-level acoustic perturbation metrics (jitter, shimmer, and HNR) and applying targeted noise-cancellation algorithms, the system trains a Random Forest ensemble model to achieve high-accuracy classification between healthy control samples and early Parkinsonian vocal patterns.
