Parkinson Disease Prediction using Handwritten Spiral and Wave Pattern

This project presents a hybrid image classification system designed to facilitate early detection of
Parkinson's disease (PD) by analyzing handwritten spiral and wave patterns. Parkinson’s disease is a progressive
neurological disorder affecting millions worldwide, and subtle motor symptoms can manifest in handwriting long
before clinical diagnosis. By leveraging machine learning and deep learning techniques, this system provides an
advanced, non-invasive screening tool that can aid in early intervention.The system employs a dual-model approach,
integrating Convolutional Neural Networks (CNN) for deep feature extraction alongside Histogram of Oriented
Gradients (HOG) with Random Forest classification for structured analysis. This fusion enhances model accuracy
beyond 85%, ensuring robust performance while maintaining computational efficiency suitable for clinical
environments. Extensive experiments confirm the model’s reliability across diverse datasets, demonstrating strong
generalization capabilities and reinforcing its potential as a cost-effective, non-invasive tool for Parkinson’s disease
screening. The project contributes to the field of computer-aided diagnosis, highlighting the advantages of hybrid AI
approaches in improving the precision and efficiency of neurological disorder detection.