Applications of a deep learning method for anti-aliasing and super-resolution in MRI

Can Zhao, M. Shao, A. Carass, Hao Li, Blake E. Dewey, Lotta M. Ellingsen, Jonghye Woo, Michael A. Guttman, Ari M Blitz, Maureen Stone, Peter A. Calabresi, H. Halperin, Jerry L. Prince

Research output: Contribution to journalArticle

Abstract

Magnetic resonance (MR) images with both high resolutions and high signal-to-noise ratios (SNRs) are desired in many clinical and research applications. However, acquiring such images takes a long time, which is both costly and susceptible to motion artifacts. Acquiring MR images with good in-plane resolution and poor through-plane resolution is a common strategy that saves imaging time, preserves SNR, and provides one viewpoint with good resolution in two directions. Unfortunately, this strategy also creates orthogonal viewpoints that have poor resolution in one direction and, for 2D MR acquisition protocols, also creates aliasing artifacts. A deep learning approach called SMORE that carries out both anti-aliasing and super-resolution on these types of acquisitions using no external atlas or exemplars has been previously reported but not extensively validated. This paper reviews the SMORE algorithm and then demonstrates its performance in four applications with the goal to demonstrate its potential for use in both research and clinical scenarios. It is first shown to improve the visualization of brain white matter lesions in FLAIR images acquired from multiple sclerosis patients. Then it is shown to improve the visualization of scarring in cardiac left ventricular remodeling after myocardial infarction. Third, its performance on multi-view images of the tongue is demonstrated and finally it is shown to improve performance in parcellation of the brain ventricular system. Both visual and selected quantitative metrics of resolution enhancement are demonstrated.

Original languageEnglish (US)
JournalMagnetic Resonance Imaging
DOIs
StatePublished - Jan 1 2019

Fingerprint

Anti-aliasing
Magnetic resonance
Magnetic resonance imaging
Magnetic Resonance Spectroscopy
Learning
Signal-To-Noise Ratio
Artifacts
Brain
Signal to noise ratio
Visualization
Ventricular Remodeling
Atlases
Tongue
Research
Multiple Sclerosis
Cicatrix
Myocardial Infarction
Imaging techniques
Deep learning
Direction compound

Keywords

  • Aliasing
  • Deep learning
  • MRI
  • Reconstruction
  • Segmentation
  • SMORE
  • Super-resolution

ASJC Scopus subject areas

  • Biophysics
  • Biomedical Engineering
  • Radiology Nuclear Medicine and imaging

Cite this

Applications of a deep learning method for anti-aliasing and super-resolution in MRI. / Zhao, Can; Shao, M.; Carass, A.; Li, Hao; Dewey, Blake E.; Ellingsen, Lotta M.; Woo, Jonghye; Guttman, Michael A.; Blitz, Ari M; Stone, Maureen; Calabresi, Peter A.; Halperin, H.; Prince, Jerry L.

In: Magnetic Resonance Imaging, 01.01.2019.

Research output: Contribution to journalArticle

Zhao, Can ; Shao, M. ; Carass, A. ; Li, Hao ; Dewey, Blake E. ; Ellingsen, Lotta M. ; Woo, Jonghye ; Guttman, Michael A. ; Blitz, Ari M ; Stone, Maureen ; Calabresi, Peter A. ; Halperin, H. ; Prince, Jerry L. / Applications of a deep learning method for anti-aliasing and super-resolution in MRI. In: Magnetic Resonance Imaging. 2019.
@article{a99de6ddd82946a095a50be23e6f7675,
title = "Applications of a deep learning method for anti-aliasing and super-resolution in MRI",
abstract = "Magnetic resonance (MR) images with both high resolutions and high signal-to-noise ratios (SNRs) are desired in many clinical and research applications. However, acquiring such images takes a long time, which is both costly and susceptible to motion artifacts. Acquiring MR images with good in-plane resolution and poor through-plane resolution is a common strategy that saves imaging time, preserves SNR, and provides one viewpoint with good resolution in two directions. Unfortunately, this strategy also creates orthogonal viewpoints that have poor resolution in one direction and, for 2D MR acquisition protocols, also creates aliasing artifacts. A deep learning approach called SMORE that carries out both anti-aliasing and super-resolution on these types of acquisitions using no external atlas or exemplars has been previously reported but not extensively validated. This paper reviews the SMORE algorithm and then demonstrates its performance in four applications with the goal to demonstrate its potential for use in both research and clinical scenarios. It is first shown to improve the visualization of brain white matter lesions in FLAIR images acquired from multiple sclerosis patients. Then it is shown to improve the visualization of scarring in cardiac left ventricular remodeling after myocardial infarction. Third, its performance on multi-view images of the tongue is demonstrated and finally it is shown to improve performance in parcellation of the brain ventricular system. Both visual and selected quantitative metrics of resolution enhancement are demonstrated.",
keywords = "Aliasing, Deep learning, MRI, Reconstruction, Segmentation, SMORE, Super-resolution",
author = "Can Zhao and M. Shao and A. Carass and Hao Li and Dewey, {Blake E.} and Ellingsen, {Lotta M.} and Jonghye Woo and Guttman, {Michael A.} and Blitz, {Ari M} and Maureen Stone and Calabresi, {Peter A.} and H. Halperin and Prince, {Jerry L.}",
year = "2019",
month = "1",
day = "1",
doi = "10.1016/j.mri.2019.05.038",
language = "English (US)",
journal = "Magnetic Resonance Imaging",
issn = "0730-725X",
publisher = "Elsevier Inc.",

}

TY - JOUR

T1 - Applications of a deep learning method for anti-aliasing and super-resolution in MRI

AU - Zhao, Can

AU - Shao, M.

AU - Carass, A.

AU - Li, Hao

AU - Dewey, Blake E.

AU - Ellingsen, Lotta M.

AU - Woo, Jonghye

AU - Guttman, Michael A.

AU - Blitz, Ari M

AU - Stone, Maureen

AU - Calabresi, Peter A.

AU - Halperin, H.

AU - Prince, Jerry L.

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Magnetic resonance (MR) images with both high resolutions and high signal-to-noise ratios (SNRs) are desired in many clinical and research applications. However, acquiring such images takes a long time, which is both costly and susceptible to motion artifacts. Acquiring MR images with good in-plane resolution and poor through-plane resolution is a common strategy that saves imaging time, preserves SNR, and provides one viewpoint with good resolution in two directions. Unfortunately, this strategy also creates orthogonal viewpoints that have poor resolution in one direction and, for 2D MR acquisition protocols, also creates aliasing artifacts. A deep learning approach called SMORE that carries out both anti-aliasing and super-resolution on these types of acquisitions using no external atlas or exemplars has been previously reported but not extensively validated. This paper reviews the SMORE algorithm and then demonstrates its performance in four applications with the goal to demonstrate its potential for use in both research and clinical scenarios. It is first shown to improve the visualization of brain white matter lesions in FLAIR images acquired from multiple sclerosis patients. Then it is shown to improve the visualization of scarring in cardiac left ventricular remodeling after myocardial infarction. Third, its performance on multi-view images of the tongue is demonstrated and finally it is shown to improve performance in parcellation of the brain ventricular system. Both visual and selected quantitative metrics of resolution enhancement are demonstrated.

AB - Magnetic resonance (MR) images with both high resolutions and high signal-to-noise ratios (SNRs) are desired in many clinical and research applications. However, acquiring such images takes a long time, which is both costly and susceptible to motion artifacts. Acquiring MR images with good in-plane resolution and poor through-plane resolution is a common strategy that saves imaging time, preserves SNR, and provides one viewpoint with good resolution in two directions. Unfortunately, this strategy also creates orthogonal viewpoints that have poor resolution in one direction and, for 2D MR acquisition protocols, also creates aliasing artifacts. A deep learning approach called SMORE that carries out both anti-aliasing and super-resolution on these types of acquisitions using no external atlas or exemplars has been previously reported but not extensively validated. This paper reviews the SMORE algorithm and then demonstrates its performance in four applications with the goal to demonstrate its potential for use in both research and clinical scenarios. It is first shown to improve the visualization of brain white matter lesions in FLAIR images acquired from multiple sclerosis patients. Then it is shown to improve the visualization of scarring in cardiac left ventricular remodeling after myocardial infarction. Third, its performance on multi-view images of the tongue is demonstrated and finally it is shown to improve performance in parcellation of the brain ventricular system. Both visual and selected quantitative metrics of resolution enhancement are demonstrated.

KW - Aliasing

KW - Deep learning

KW - MRI

KW - Reconstruction

KW - Segmentation

KW - SMORE

KW - Super-resolution

UR - http://www.scopus.com/inward/record.url?scp=85069617360&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85069617360&partnerID=8YFLogxK

U2 - 10.1016/j.mri.2019.05.038

DO - 10.1016/j.mri.2019.05.038

M3 - Article

C2 - 31247254

AN - SCOPUS:85069617360

JO - Magnetic Resonance Imaging

JF - Magnetic Resonance Imaging

SN - 0730-725X

ER -