Design of an Efficient Forecast Model for Too Early Parkinson’s Disease Detection

Abstract

Parkinson's Disease (PD) presents a formidable challenge in diagnosis due to the absence of definitive clinical tests, especially in its early stages. This project addresses the critical need for an effective and non-invasive methodology for early PD detection. This project leverages the power of deep learning, specifically Convolutional Neural Networks (CNNs), to analyze handwriting patterns for PD detection. Deep learning has shown exceptional accuracy in classification tasks, and its application in medical fields, interpreting data like X-rays and MRI scans, has proven advantageous. The project focuses on optimizing accuracy by extracting features from handwriting, training machine learning models, and comparing their performance against traditional methods.In addition to leveraging VGG19, InceptionV3, and ResNet50 for feature extraction, this project extends its capabilities by incorporating the Xception algorithm for enhanced feature extraction. The classification process is fortified through the implementation of a robust voting classifier. To facilitate user testing and practical application, a Flask framework with SQLite ssssintegration has been developed. This project not only broadens the scope of the project but also ensures a user-friendly interface for real-world testing and validation in clinical settings.