Automated Early Detection of Parkinson's Disease Through Augmented Handwritten Patterns

Emerging eHealth tools including wearable technology and telemedicine are becoming increasingly popular. This advancement revolutionizes the early detection and management of diseases such as Parkinson Disease (PD). Due to the loss of dopamine-producing neurons, PD is a progressive central nervous system condition that affects fundamental motor abilities, with hand tremors often being the first symptom. Despite the importance of early diagnosis, there is currently no conclusive test for PD. In our work, we consider a diagnostic tool for the early detection and diagnosis of PD by analyzing hand-drawn spiral and wave patterns. Hence, we aim in our model to automate the PD $\mathbf{s}$ early diagnosis. This innovation opens up a realm of possibilities, particularly in enabling remote monitoring and timely intervention. In this paper, we use two deep learning models, Inception V3 and Xception, to predict the risk of PD based on the pattern of the drawings. The models were trained on augmented data consisting of more than 4500 images. Our diagnostic tool achieves an accuracy rate of 97% which ultimately improves the lives of countless individuals battling this challenging condition.