DeepResp: Deep learning solution for respiration-induced B0 fluctuation artifacts in multi-slice GRE

Respiration-induced B\(_0\) fluctuation corrupts MRI images by inducing phase errors in k-space. A few approaches such as navigator have been proposed to correct for the artifacts at the expense of sequence modification. In this study, a new deep learning method, which is referred to as DeepResp, is...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2020-07
Main Authors: An, Hongjun, Shin, Hyeong-Geol, Ji, Sooyoen, Jung, Woojin, Oh, Sehong, Shin, Dongmyung, Park, Juhyung, Lee, Jongho
Format: Article
Language:English
Subjects:
Artificial neural networks
Breathing
Computer simulation
Deep learning
Machine learning
Magnetic resonance imaging
Neural networks
Respiration
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Respiration-induced B\(_0\) fluctuation corrupts MRI images by inducing phase errors in k-space. A few approaches such as navigator have been proposed to correct for the artifacts at the expense of sequence modification. In this study, a new deep learning method, which is referred to as DeepResp, is proposed for reducing the respiration-artifacts in multi-slice gradient echo (GRE) images. DeepResp is designed to extract the respiration-induced phase errors from a complex image using deep neural networks. Then, the network-generated phase errors are applied to the k-space data, creating an artifact-corrected image. For network training, the computer-simulated images were generated using artifact-free images and respiration data. When evaluated, both simulated images and in-vivo images of two different breathing conditions (deep breathing and natural breathing) show improvements (simulation: normalized root-mean-square error (NRMSE) from 7.8% to 1.3%; structural similarity (SSIM) from 0.88 to 0.99; ghost-to-signal-ratio (GSR) from 7.9% to 0.6%; deep breathing: NRMSE from 13.9% to 5.8%; SSIM from 0.86 to 0.95; GSR 20.2% to 5.7%; natural breathing: NRMSE from 5.2% to 4.0%; SSIM from 0.94 to 0.97; GSR 5.7% to 2.8%). Our approach does not require any modification of the sequence or additional hardware, and may therefore find useful applications. Furthermore, the deep neural networks extract respiration-induced phase errors, which is more interpretable and reliable than results of end-to-end trained networks.
ISSN:2331-8422
DOI:10.48550/arxiv.2007.09597