Cardiac mr motion artefact correction from k-space using deep learning-based reconstruction

Ilkay Oksuz; James Clough; Aurelien Bustin; Gastao Cruz; Claudia Prieto; Rene Botnar; Daniel Rueckert; Julia A. Schnabel; Andrew P. King

doi:10.1007/978-3-030-00129-2_3

Cardiac mr motion artefact correction from k-space using deep learning-based reconstruction

Ilkay Oksuz^*
, James Clough
, Aurelien Bustin
, Gastao Cruz
, Claudia Prieto
, Rene Botnar
, Daniel Rueckert
, Julia A. Schnabel
, Andrew P. King

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

22 Citations (Scopus)

Abstract

Incorrect ECG gating of cardiac magnetic resonance (CMR) acquisitions can lead to artefacts, which hampers the accuracy of diagnostic imaging. Therefore, there is a need for robust reconstruction methods to ensure high image quality. In this paper, we propose a method to automatically correct motion-related artefacts in CMR acquisitions during reconstruction from k-space data. Our method is based on the Automap reconstruction method, which directly reconstructs high quality MR images from k-space using deep learning. Our main methodological contribution is the addition of an adversarial element to this architecture, in which the quality of image reconstruction (the generator) is increased by using a discriminator. We train the reconstruction network to automatically correct for motion-related artefacts using synthetically corrupted CMR k-space data and uncorrupted reconstructed images. Using 25000 images from the UK Biobank dataset we achieve good image quality in the presence of synthetic motion artefacts, but some structural information was lost. We quantitatively compare our method to a standard inverse Fourier reconstruction. In addition, we qualitatively evaluate the proposed technique using k-space data containing real motion artefacts.

Original language	English
Title of host publication	Machine Learning for Medical Image Reconstruction - First International Workshop, MLMIR 2018, Held in Conjunction with MICCAI 2018, Proceedings
Editors	Florian Knoll, Andreas Maier, Daniel Rueckert
Publisher	Springer Verlag
Pages	21-29
Number of pages	9
ISBN (Print)	9783030001285
DOIs	https://doi.org/10.1007/978-3-030-00129-2_3
Publication status	Published - 2018
Externally published	Yes
Event	1st Workshop on Machine Learning for Medical Image Reconstruction, MLMIR 2018 Held in Conjunction with 21st Medical Image Computing and Computer Assisted Intervention, MICCAI 2018 - Granada, Spain Duration: 16 Sept 2018 → 16 Sept 2018

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11074 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	1st Workshop on Machine Learning for Medical Image Reconstruction, MLMIR 2018 Held in Conjunction with 21st Medical Image Computing and Computer Assisted Intervention, MICCAI 2018
Country/Territory	Spain
City	Granada
Period	16/09/18 → 16/09/18

Bibliographical note

Publisher Copyright:
© 2018, Springer Nature Switzerland AG.

Funding

Acknowledgments. This work was supported by an EPSRC programme Grant (EP/P001009/1) and the Wellcome EPSRC Centre for Medical Engineering at the School of Biomedical Engineering and Imaging Sciences, King’s College London (WT 203148/Z/16/Z). This research has been conducted using the UK Biobank Resource under Application Number 17806. The GPU used in this research was generously donated by the NVIDIA Corporation.

Funders	Funder number
Wellcome EPSRC Centre for Medical Engineering at the School of Biomedical Engineering and Imaging Sciences
King’s College London	WT 203148/Z/16/Z
Engineering and Physical Sciences Research Council	EP/P001009/1, EP/N026993/1, EP/M000133/1

Keywords

Automap
Cardiac MR
Deep learning
Image artefacts
Image quality
Image reconstruction
UK Biobank

Access to Document

10.1007/978-3-030-00129-2_3

Cite this

Oksuz, I., Clough, J., Bustin, A., Cruz, G., Prieto, C., Botnar, R., Rueckert, D., Schnabel, J. A., & King, A. P. (2018). Cardiac mr motion artefact correction from k-space using deep learning-based reconstruction. In F. Knoll, A. Maier, & D. Rueckert (Eds.), Machine Learning for Medical Image Reconstruction - First International Workshop, MLMIR 2018, Held in Conjunction with MICCAI 2018, Proceedings (pp. 21-29). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11074 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-00129-2_3

Oksuz, Ilkay ; Clough, James ; Bustin, Aurelien et al. / Cardiac mr motion artefact correction from k-space using deep learning-based reconstruction. Machine Learning for Medical Image Reconstruction - First International Workshop, MLMIR 2018, Held in Conjunction with MICCAI 2018, Proceedings. editor / Florian Knoll ; Andreas Maier ; Daniel Rueckert. Springer Verlag, 2018. pp. 21-29 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{3d5b449649d34b3cbf67d664877a91b0,

title = "Cardiac mr motion artefact correction from k-space using deep learning-based reconstruction",

abstract = "Incorrect ECG gating of cardiac magnetic resonance (CMR) acquisitions can lead to artefacts, which hampers the accuracy of diagnostic imaging. Therefore, there is a need for robust reconstruction methods to ensure high image quality. In this paper, we propose a method to automatically correct motion-related artefacts in CMR acquisitions during reconstruction from k-space data. Our method is based on the Automap reconstruction method, which directly reconstructs high quality MR images from k-space using deep learning. Our main methodological contribution is the addition of an adversarial element to this architecture, in which the quality of image reconstruction (the generator) is increased by using a discriminator. We train the reconstruction network to automatically correct for motion-related artefacts using synthetically corrupted CMR k-space data and uncorrupted reconstructed images. Using 25000 images from the UK Biobank dataset we achieve good image quality in the presence of synthetic motion artefacts, but some structural information was lost. We quantitatively compare our method to a standard inverse Fourier reconstruction. In addition, we qualitatively evaluate the proposed technique using k-space data containing real motion artefacts.",

keywords = "Automap, Cardiac MR, Deep learning, Image artefacts, Image quality, Image reconstruction, UK Biobank",

author = "Ilkay Oksuz and James Clough and Aurelien Bustin and Gastao Cruz and Claudia Prieto and Rene Botnar and Daniel Rueckert and Schnabel, \{Julia A.\} and King, \{Andrew P.\}",

note = "Publisher Copyright: {\textcopyright} 2018, Springer Nature Switzerland AG.; 1st Workshop on Machine Learning for Medical Image Reconstruction, MLMIR 2018 Held in Conjunction with 21st Medical Image Computing and Computer Assisted Intervention, MICCAI 2018 ; Conference date: 16-09-2018 Through 16-09-2018",

year = "2018",

doi = "10.1007/978-3-030-00129-2\_3",

language = "English",

isbn = "9783030001285",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "21--29",

editor = "Florian Knoll and Andreas Maier and Daniel Rueckert",

booktitle = "Machine Learning for Medical Image Reconstruction - First International Workshop, MLMIR 2018, Held in Conjunction with MICCAI 2018, Proceedings",

address = "Germany",

}

Oksuz, I, Clough, J, Bustin, A, Cruz, G, Prieto, C, Botnar, R, Rueckert, D, Schnabel, JA & King, AP 2018, Cardiac mr motion artefact correction from k-space using deep learning-based reconstruction. in F Knoll, A Maier & D Rueckert (eds), Machine Learning for Medical Image Reconstruction - First International Workshop, MLMIR 2018, Held in Conjunction with MICCAI 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11074 LNCS, Springer Verlag, pp. 21-29, 1st Workshop on Machine Learning for Medical Image Reconstruction, MLMIR 2018 Held in Conjunction with 21st Medical Image Computing and Computer Assisted Intervention, MICCAI 2018, Granada, Spain, 16/09/18. https://doi.org/10.1007/978-3-030-00129-2_3

Cardiac mr motion artefact correction from k-space using deep learning-based reconstruction. / Oksuz, Ilkay; Clough, James; Bustin, Aurelien et al.
Machine Learning for Medical Image Reconstruction - First International Workshop, MLMIR 2018, Held in Conjunction with MICCAI 2018, Proceedings. ed. / Florian Knoll; Andreas Maier; Daniel Rueckert. Springer Verlag, 2018. p. 21-29 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11074 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Cardiac mr motion artefact correction from k-space using deep learning-based reconstruction

AU - Oksuz, Ilkay

AU - Clough, James

AU - Bustin, Aurelien

AU - Cruz, Gastao

AU - Prieto, Claudia

AU - Botnar, Rene

AU - Rueckert, Daniel

AU - Schnabel, Julia A.

AU - King, Andrew P.

PY - 2018

Y1 - 2018

N2 - Incorrect ECG gating of cardiac magnetic resonance (CMR) acquisitions can lead to artefacts, which hampers the accuracy of diagnostic imaging. Therefore, there is a need for robust reconstruction methods to ensure high image quality. In this paper, we propose a method to automatically correct motion-related artefacts in CMR acquisitions during reconstruction from k-space data. Our method is based on the Automap reconstruction method, which directly reconstructs high quality MR images from k-space using deep learning. Our main methodological contribution is the addition of an adversarial element to this architecture, in which the quality of image reconstruction (the generator) is increased by using a discriminator. We train the reconstruction network to automatically correct for motion-related artefacts using synthetically corrupted CMR k-space data and uncorrupted reconstructed images. Using 25000 images from the UK Biobank dataset we achieve good image quality in the presence of synthetic motion artefacts, but some structural information was lost. We quantitatively compare our method to a standard inverse Fourier reconstruction. In addition, we qualitatively evaluate the proposed technique using k-space data containing real motion artefacts.

AB - Incorrect ECG gating of cardiac magnetic resonance (CMR) acquisitions can lead to artefacts, which hampers the accuracy of diagnostic imaging. Therefore, there is a need for robust reconstruction methods to ensure high image quality. In this paper, we propose a method to automatically correct motion-related artefacts in CMR acquisitions during reconstruction from k-space data. Our method is based on the Automap reconstruction method, which directly reconstructs high quality MR images from k-space using deep learning. Our main methodological contribution is the addition of an adversarial element to this architecture, in which the quality of image reconstruction (the generator) is increased by using a discriminator. We train the reconstruction network to automatically correct for motion-related artefacts using synthetically corrupted CMR k-space data and uncorrupted reconstructed images. Using 25000 images from the UK Biobank dataset we achieve good image quality in the presence of synthetic motion artefacts, but some structural information was lost. We quantitatively compare our method to a standard inverse Fourier reconstruction. In addition, we qualitatively evaluate the proposed technique using k-space data containing real motion artefacts.

KW - Automap

KW - Cardiac MR

KW - Deep learning

KW - Image artefacts

KW - Image quality

KW - Image reconstruction

KW - UK Biobank

UR - https://www.scopus.com/pages/publications/85053935756

U2 - 10.1007/978-3-030-00129-2_3

DO - 10.1007/978-3-030-00129-2_3

M3 - Conference contribution

AN - SCOPUS:85053935756

SN - 9783030001285

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 21

EP - 29

BT - Machine Learning for Medical Image Reconstruction - First International Workshop, MLMIR 2018, Held in Conjunction with MICCAI 2018, Proceedings

A2 - Knoll, Florian

A2 - Maier, Andreas

A2 - Rueckert, Daniel

PB - Springer Verlag

T2 - 1st Workshop on Machine Learning for Medical Image Reconstruction, MLMIR 2018 Held in Conjunction with 21st Medical Image Computing and Computer Assisted Intervention, MICCAI 2018

Y2 - 16 September 2018 through 16 September 2018

ER -

Oksuz I, Clough J, Bustin A, Cruz G, Prieto C, Botnar R et al. Cardiac mr motion artefact correction from k-space using deep learning-based reconstruction. In Knoll F, Maier A, Rueckert D, editors, Machine Learning for Medical Image Reconstruction - First International Workshop, MLMIR 2018, Held in Conjunction with MICCAI 2018, Proceedings. Springer Verlag. 2018. p. 21-29. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-00129-2_3

Cardiac mr motion artefact correction from k-space using deep learning-based reconstruction

Abstract

Publication series

Conference

Bibliographical note

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this