Optimized Parkinson’s Disease Classification on Hand Drawing Images using Transfer Learning Models: An Exploratory Analysis

Parkinson's disease (PD) is a progressive neurological disorder that impairs motor functions, leading to symptoms such as tremor, muscular rigidity, cognitive instability and other related health disorders. Early diagnosis is essential for effective therapeutic management. The advancements in artificial Intelligence offer innovative and Novel solutions for automated diagnosis of this health disorder, since traditional diagnostic methods are often subjective, time-consuming, and reliant on clinical expertise. In recent years, machine learning (ML) and deep learning (DL) have emerged as promising tools for automating disease detection and improving diagnostic accuracy. This study proposes a deep learning-based approach for diagnosing Parkinson's disease using hand-drawn images from two publicly available datasets: the Parkinson's Drawing Dataset and the NewHandPD Dataset. Five state-of-the-art deep learning models—InceptionV3, Xception, ResNet152, VGG19, and EfficientNet were employed for feature extraction and image classification. The models categorize input images as either Parkinson’s or non-Parkinson’s, with an evaluation, based on performance metrics including accuracy, precision, recall, and F1-Score. The goal is to enhance early PD detection through robust and scalable automated analysis methods.