We propose a novel deep learning-based approach for accelerated 4D Flow MRI by reducing artifact in complex image domain from undersampled k-space. A deep 2D residual attention network is proposed which is trained independently for three velocity-sensitive encoding directions to learn the mapping of complex image from zero-filled reconstruction to complex image from fully sampled k-space. Network is trained and tested on 4D flow MRI data of aortic valvular flow in 18 human subjects. Proposed method outperforms state of the art TV regularized reconstruction method and deep learning reconstruction approach by U-net.
This abstract and the presentation materials are available to members only; a login is required.