Correction of Motion Artifacts Using a Multiscale Fully Convolutional Neural Network

Abstract

Retrospective correction of motion artifacts using a multiscale fully convolutional network is promising and may mitigate the substantial motion-related problems in the clinical MRI workflow.