Brain Tumour Diagnosis with Lightweight Federated Learning using Identically Distributed Images

The innovative method was developed by bringing federated learning methods to the medical field. In this study, the author refined federated learning models with data sets containing photographs of brain tumors. The cancer research community faces a huge problem in trying to predict the development of brain tumors. This paper proposes a streamlined framework for using federated deep learning to classify brain tumor diseases while keeping private information safe. When validating federated deep learning models, this system leverages a dispersed client-server architecture to safeguard client data and IID. The system was tested in both conventional and federated classroom settings with a small and large number of users. After evaluating several pre-trained models, we chose AlexNet as our benchmark due to its high 98.53 percent accuracy. Federal learning (FL) was then used to assess IID datasets. The framework is the best choice for detecting and categorizing brain tumors early because of its improved capabilities.