Automated Parkinson's disease Detection from Images Using Deep Transfer Learning and Optimization

Abstract

The diagnosis and treatment of Parkinson's disease (PD) is critical to effectively managing this progressive neurological disorder, which significantly affects motor and non-motor functions. This study presents a deep transfer learning-based algorithm for PD detection. The features are extracted from handwritten image datasets using pre-trained convolutional neural networks such as ResNet50, VGG19, and Inception-V3. To achieve precise classification, a hybrid classification framework that combines a genetic algorithm-optimized k-nearest neighbour (KNN) classifier with a support vector machine (SVM) is implemented. The proposed model offers a reliable, scalable, and efficient solution for diagnosing Parkinson's disease. The experimental results demonstrate the model's state-of-the-art accuracy. AI-driven methodologies are being used in this research to advance automated medical diagnostics, reduce diagnostic delays, and improve patient outcomes.