EHM-NET: A NOVEL ENSEMBLE DEEP LEARNING FRAMEWORK FOR EARLY DETECTION OF PARKINSON’S DISEASE USING HANDWRITING AND MRI MODALITIES

Parkinson’s Disease (PD) is a progressive neurodegenerative disorder affecting millions worldwide. Early and accurate detection of PD remains a critical challenge. In this work, we propose a novel deep learning ensemble framework that simultaneously analyzes multimodal data – specifically, handwritten spiral/wave drawings and brain MRI scans – to detect PD. We introduce a two-stage ensemble architecture combining multiple convolutional neural networks (CNNs) and a meta-learner to integrate complementary features. Our method leverages the publicly available Parkinson’s Drawings dataset and the PPMI MRI dataset. In extensive experiments, our proposed model achieves classification accuracies exceeding those reported in recent studies. Table 1 compares our performance to state-of-the-art methods, showing clear improvements. Detailed analysis and ablation studies show the effectiveness of our approach. These results suggest that multimodal deep ensembles can significantly enhance early PD diagnosis.