Analysis of deep complex‐valued convolutional neural networks for MRI reconstruction and phase‐focused applications

Abstract

Complex-valued CNNs can enable superior accelerated MRI reconstruction and phase-based applications such as fat-water separation, and flow quantification compared to real-valued convolutional neural networks.