PARKINSON’S DISEASE DETECTION USING DEEP LEARNING

COMPUTER ENGINEERING

Parkinson's disease, the second most prevalent degenerative disorder stemming from the loss of dopamine-producing neurons, primarily affects the substantia nigra region, resulting in a deficiency of dopamine in the striatum, which is a key characteristic of the disease. Clinical diagnosis encompasses a spectrum of both motor and non-motor symptoms. Magnetic Resonance Imaging (MRI) serves as a tool to visualize structural changes in the brain due to dopamine deficiency in individuals with Parkinson's disease. This study employs deep learning techniques, specifically the Convolutional Neural Network architecture Resnet, to classify MRI images of healthy controls and Parkinson's disease patients. By training the network with transfer learning, the model can accurately distinguish between the two groups, achieving an accuracy of 91.9%. This approach demonstrates the potential for deep learning models to aid clinicians in Parkinson's disease diagnosis, offering an objective and improved method for patient classification in the future.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.