Micrographia-based parkinson's disease detection using Deep Learning

Parkinson's Disease (PD) is a disorder of the nervous system that is chronic and progressive, and affects millions of people around the globe. PD often manifests through symptoms like tremors or shaking, slowness of movement (Bradykinesia), freezing of gait, impaired posture, muscle stiffness, and others. The key focus is the early diagnosis of PD symptoms, which, if handled in their initial stage, can improve the quality of life for the patients. The fine motor control of PD-affected persons, particularly handwriting (Micrographia), can be used for PD diagnosis in patients. Deep Learning (DL) approaches, a subfield of machine learning research represent a useful tool for unsupervised feature learning because they employ a succession of layers, each of which is responsible for extracting different sorts of data. This research work utilises Convolutional Neural Networks (a deep learning algorithm) and focuses on Micrographia as the main diagnostic feature of PD. This work has achieved two main goals, i.e. utilizing CNN for PD diagnosis by learning features from handwriting, thereby improving, and assisting in PD detection, and enhancing overall diagnostic accuracy. The proposed system has achieved the following metrics: Accuracy of 96.67%, Precision of 96.67%, and Recall of 96.67%.