Compressed Sensing and Parallel MRI using Deep Residual Learning

Abstract

A deep residual learning algorithm is proposed to reconstruct MR images from highly down-sampled k-space data. After formulating a compressed sensing problem as a residual regression problem, a deep convolutional neural network (CNN) was designed to learn the aliasing artifacts. The residual learning algorithm took only 30-40ms with significantly better reconstruction performance compared to GRAPPA and the state-of-the-art compressed sensing algorithm, ALOHA.