Enhanced Parkinson’s Disease Diagnosis via MRI Analysis: Integrating Deep Features From DenseNet201 With Neural Network Techniques

Abstract

Parkinson’s disease (PD) is a neurodegenerative disorder that affects millions of people worldwide, necessitating accurate and timely diagnostic methods for effective disease management. This study proposes a novel approach for PD detection using deep features extracted from magnetic resonance imaging (MRI) scans, employing a pattern recognition neural network architecture based on DenseNet201. The developed model demonstrated exceptional performance, achieving validation and test accuracies of 99.4% and 99.2%, respectively, indicating its robustness and efficacy in distinguishing between patients with PD and healthy individuals. Furthermore, the model achieved a precision of 99.5%, recall of 99.3%, and an F1 score of 99.4%. For the test set, the accuracy was 99.2%, with precision at 99.3%, recall at 99.1%, and an F1 score of 99.2%. The rapid convergence of the model during training further underscores its efficiency in learning discriminative features from MRI images. These findings underscore the promising role of deep learning techniques, particularly convolutional neural networks (CNNs), in medical image analysis for disease diagnosis. The proposed approach holds significant potential for assisting clinicians in early PD diagnosis and personalized treatment planning, ultimately improving patient outcomes and quality of life.