CNN –based Spiral drawing classifier to detect Early-Stage Parkinson’s Disease

Abstract

Abstract Parkinson’s disease is a neurological condition wherein one experiences a loss in control of motor movements. The disease usually infects gradually but worsens over time. Parkinson’s often starts with a tremor in one hand. Slow movements, stiffness in the body, and loss of balance are some of the major symptoms of the disease. Damaged nerve cells present in the brain cause drops in the levels of dopamine in the body. It is very difficult to detect Early-Stage Parkinson’s Disease. Researchers have now been able to detect Parkinson’s disease at an early stage (Perhaps at Stage 2 or Stage 3), but there have been no significant advancements in the detection of Parkinson’s at the initial stage. Analyzing the handwriting styles can be one of the initial symptoms of the disease. The dataset given by NIATS of the Federal University of Uberlandia was used in our research. In this study, we present a CNN model that can distinguish between spiral drawings made by patients with Parkinson’s disease and those made by people not suffering from Parkinson’s disease with high accuracy.