Reconstruction of sparsely sampled Magnetic Resonance Imaging measurements with a convolutional neural network

Jasper Schoormans, Qinwei Zhang, Bram Coolen, Gustav Strijkers, Aart Nederveen

Abstract:

Compressed Sensing accelerated Magnetic Resonance Imaging (MRI) suffers from long image reconstruction times, due to the need for solving ill-posed minimizations. This limits the clinical use of accelerated MRI techniques. We have trained a
neural network to decode accelerated, undersampled MR acquisitions, eliminating the need for reconstruction algorithms.

Author Affiliation: Academic Medical Center Amsterdam