Brain Tumor Classification and Diagnosis using Federated Learning with Transfer Learning

Important yet difficult in medical imaging is the categorization of brain tumors(BTs) using magnetic-resonance-imaging(MRI) scans. The intricacy of tumor form and imaging variations frequently impede accurate diagnosis, which is critical for efficient treatment planning. Accurate and automated tumor categorization is generally impossible with these methods due to their lack of scalability and stability. We offer a deep learning model based on federated learning that uses CNNs for automatically and accurately classify brain tumors in order to overcome these problems. Using a federated learning and transfer learning methodology, this novel method stresses the usage of a modified VGG16 that tuned for brain MRI-images. By including pre-trained convolutional-neural-network(CNN), this model architecture takes advantage of the transfer learning technique. This improves its capacity of effectively classify brain tumors by drawing on insights obtained from large and varied datasets. Overall, the model achieves 98.5% accuracy, demonstrating its effectiveness in diagnosing different types of tumors.