MRI is short for magnetic resonance imaging, which uses a strong magnetic field and radio waves in order to non-invasively create detailed images of organs and tissues inside the body. The problem is that MRI scans are notoriously slow, and a single scan can take up to an hour to complete. Our project aims to address this by building a web application for hosting federated learning for reconstructing MRI scans from undersampled data. Federated learning refers to the process of aggregating many locally trained models into a single global model that can be shared with everyone. This way, researchers and hospitals can collaborate to train models for MRI reconstruction without sharing private patient information with others.

Team Members

Sugam Arora
Ammar Fatehi
Mervin John
Emran Khan
Jason Wang
Bijan Ziaie