NIH - LOVIR - Learning an Optimized Variational Network for Medical Image Reconstruction

The overarching goal of the proposal is to use a novel deep learning approach for MR image reconstruction of accelerated protocols that is able to learn the characteristic appearance of clinical imaging datasets, as well as suppression of artifacts that arise during data acquisition. The results of this proposal will allow acceleration of current clinical protocols and near-real-time reconstruction while preserving the diagnostic content and natural appearance of the images.