Efficient Alzheimer's Disease Classification Using Semi-Supervised Federated Learning

Alzheimer's disease has become one of the most dangerous disease that can endanger patients' lives if left undiagnosed because there is no known cause for its effects. Medical professionals and researchers are working to solve the problem, and they might be successful soon. In order to solve this problem, new methods must be implemented. This research offers a new method for using deep machine learning to classify Alzheimer's disease. By using the ADNI MRI dataset, we use a convolutional neural network (CNNs) in conjunction with semisupervised federated learning with secured aggregation to classify Alzheimer's disease. Additionally, this work demonstrates how we bind MRI scans for various goals, utilizing decentralized training to learn the model for both labeled and unlabeled training images.Through the use of learning techniques, we create a system that capable of classifying, detecting, and diagnosing Alzheimer's disease with an excellent 99.3 % testing accuracy. As soon as symptoms start to show up this enables our physicians to identify the patients' condition at various stages and physicians should take necessary initiative for appropriate disease diagnosis.Our research also giving us important information on how our deep learning models make decisions. This transparency promotes the AI-driven disease classification system's usefulness in medical by enhancing its credibility.

Ҳали таржима қилинмаган