Decentralized Brain Tumor Detection: Leveraging Federated Learning & CNN for Enhanced Accuracy and Privacy-A Review
DOI : 10.46335/IJIES.2024.9.3.6
Authors : Mr.Jagdish F. Pimple, Dr.Mohit Agarwal, Mr. Sameer Kedar, Ms.Isha Bhomle, Ms.Nishita Thakare, Ms.Pratiksha Bowade
Abstract: This study introduces innovative brain tumor diagnostic techniques utilizing federated learning and machine learning to develop a decentralized model that safeguards patient data privacy. Our methodology includes the collection and preprocessing of MRI datasets, implementation of a Convolutional Neural Network (CNN) core model, and creation of an intuitive Flask-based web application for seamless MRI image upload. Rigorous evaluations are conducted to ensure accurate diagnostic outcomes, with a strong emphasis on privacy through data encryption and adherence to healthcare regulations. This approach notably enhances tumor recognition accuracy while preserving confidentiality. The system effectively detects pituitary tumors, gliomas, meningiomas, and normal MRI scans, representing a significant advancement in medical diagnosis.
