Abstract

Parkinson’s disease (PD) is a prevalent neurological condition affecting a significant number of people worldwide. Parkinson’s disease (PD) is characterized by physical indications, including rigidity, shaking, and postural instability. Timely and precise diagnosis is crucial for effective intervention as well as disease management. However, Diagnosing Parkinson's disease (PD) and keeping track of its progression can be costly and inconvenient. Early-stage symptoms often manifest as handwriting disorders and changes in voice frequency. Detecting the disease based on handwriting could prove to be a superior technique, particularly utilizing spiral drawings. Spiral drawings are easily obtainable. We employ deep learning techniques with a primary focus on improving detection accuracy, particularly utilizing Feed forward Neural Networks (FNN) with Multilayer Perceptron (MLP) Architecture and HOG (Histogram of Oriented Gradients) Algorithm. Additionally, we have developed an interface that specifically requests input in the form of an individual's spiral drawing image.