Generalized framework using Federated Learning for brain tumor detection from medical images

This paper presents federated learning to identify brain tumors based on MRI scans. A federated learning approach enables a model to be trained on decentralized data sources without requiring direct data exchange. In the field of medical imaging, where privacy rules restrict data exchange, this technique is very beneficial. The study investigates the categorization of brain tumors using EfficientNet, a convolutional neural network design. A federated learning configuration with twenty clients was designed to evaluate model performance. The results show that the federated learning approach using EfficientNet achieved an accuracy of 97.56% in detecting brain tumors. This research found that federated learning is valuable in brain tumor detection systems.