Deep learning using K-space based data augmentation for automated cardiac MR motion artefact detection

Ilkay Oksuz; Bram Ruijsink; Esther Puyol-Antón; Aurelien Bustin; Gastao Cruz; Claudia Prieto; Daniel Rueckert; Julia A. Schnabel; Andrew P. King

doi:10.1007/978-3-030-00928-1_29

Deep learning using K-space based data augmentation for automated cardiac MR motion artefact detection

Ilkay Oksuz^*
, Bram Ruijsink
, Esther Puyol-Antón
, Aurelien Bustin
, Gastao Cruz
, Claudia Prieto
, 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

19 Citations (Scopus)

Abstract

Quality assessment of medical images is essential for complete automation of image processing pipelines. For large population studies such as the UK Biobank, artefacts such as those caused by heart motion are problematic and manual identification is tedious and time-consuming. Therefore, there is an urgent need for automatic image quality assessment techniques. In this paper, we propose a method to automatically detect the presence of motion-related artefacts in cardiac magnetic resonance (CMR) images. As this is a highly imbalanced classification problem (due to the high number of good quality images compared to the low number of images with motion artefacts), we propose a novel k-space based training data augmentation approach in order to address this problem. Our method is based on 3D spatio-temporal Convolutional Neural Networks, and is able to detect 2D+time short axis images with motion artefacts in less than 1 ms. We test our algorithm on a subset of the UK Biobank dataset consisting of 3465 CMR images and achieve not only high accuracy in detection of motion artefacts, but also high precision and recall. We compare our approach to a range of state-of-the-art quality assessment methods.

Original language	English
Title of host publication	Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings
Editors	Julia A. Schnabel, Christos Davatzikos, Carlos Alberola-López, Gabor Fichtinger, Alejandro F. Frangi
Publisher	Springer Verlag
Pages	250-258
Number of pages	9
ISBN (Print)	9783030009274
DOIs	https://doi.org/10.1007/978-3-030-00928-1_29
Publication status	Published - 2018
Externally published	Yes
Event	21st International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2018 - Granada, Spain Duration: 16 Sept 2018 → 20 Sept 2018

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11070 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	21st International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2018
Country/Territory	Spain
City	Granada
Period	16/09/18 → 20/09/18

Bibliographical note

Publisher Copyright:
© Springer Nature Switzerland AG 2018.

Funding

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. 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
Wellcome Trust
Engineering and Physical Sciences Research Council	EP/P001009/1, EP/N026993/1, EP/M000133/1

Keywords

Cardiac MR
Convolutional Neural Networks
Image quality assessment
Motion artefacts
UK Biobank

Access to Document

10.1007/978-3-030-00928-1_29

Cite this

Oksuz, I., Ruijsink, B., Puyol-Antón, E., Bustin, A., Cruz, G., Prieto, C., Rueckert, D., Schnabel, J. A., & King, A. P. (2018). Deep learning using K-space based data augmentation for automated cardiac MR motion artefact detection. In J. A. Schnabel, C. Davatzikos, C. Alberola-López, G. Fichtinger, & A. F. Frangi (Eds.), Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings (pp. 250-258). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11070 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-00928-1_29

Oksuz, Ilkay ; Ruijsink, Bram ; Puyol-Antón, Esther et al. / Deep learning using K-space based data augmentation for automated cardiac MR motion artefact detection. Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings. editor / Julia A. Schnabel ; Christos Davatzikos ; Carlos Alberola-López ; Gabor Fichtinger ; Alejandro F. Frangi. Springer Verlag, 2018. pp. 250-258 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{4d456bf730524e828bdca12b3dafe9d7,

title = "Deep learning using K-space based data augmentation for automated cardiac MR motion artefact detection",

abstract = "Quality assessment of medical images is essential for complete automation of image processing pipelines. For large population studies such as the UK Biobank, artefacts such as those caused by heart motion are problematic and manual identification is tedious and time-consuming. Therefore, there is an urgent need for automatic image quality assessment techniques. In this paper, we propose a method to automatically detect the presence of motion-related artefacts in cardiac magnetic resonance (CMR) images. As this is a highly imbalanced classification problem (due to the high number of good quality images compared to the low number of images with motion artefacts), we propose a novel k-space based training data augmentation approach in order to address this problem. Our method is based on 3D spatio-temporal Convolutional Neural Networks, and is able to detect 2D+time short axis images with motion artefacts in less than 1 ms. We test our algorithm on a subset of the UK Biobank dataset consisting of 3465 CMR images and achieve not only high accuracy in detection of motion artefacts, but also high precision and recall. We compare our approach to a range of state-of-the-art quality assessment methods.",

keywords = "Cardiac MR, Convolutional Neural Networks, Image quality assessment, Motion artefacts, UK Biobank",

author = "Ilkay Oksuz and Bram Ruijsink and Esther Puyol-Ant{\'o}n and Aurelien Bustin and Gastao Cruz and Claudia Prieto and Daniel Rueckert and Schnabel, \{Julia A.\} and King, \{Andrew P.\}",

note = "Publisher Copyright: {\textcopyright} Springer Nature Switzerland AG 2018.; 21st International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2018 ; Conference date: 16-09-2018 Through 20-09-2018",

year = "2018",

doi = "10.1007/978-3-030-00928-1\_29",

language = "English",

isbn = "9783030009274",

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

publisher = "Springer Verlag",

pages = "250--258",

editor = "Schnabel, \{Julia A.\} and Christos Davatzikos and Carlos Alberola-L{\'o}pez and Gabor Fichtinger and Frangi, \{Alejandro F.\}",

booktitle = "Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings",

address = "Germany",

}

Oksuz, I, Ruijsink, B, Puyol-Antón, E, Bustin, A, Cruz, G, Prieto, C, Rueckert, D, Schnabel, JA & King, AP 2018, Deep learning using K-space based data augmentation for automated cardiac MR motion artefact detection. in JA Schnabel, C Davatzikos, C Alberola-López, G Fichtinger & AF Frangi (eds), Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11070 LNCS, Springer Verlag, pp. 250-258, 21st International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2018, Granada, Spain, 16/09/18. https://doi.org/10.1007/978-3-030-00928-1_29

Deep learning using K-space based data augmentation for automated cardiac MR motion artefact detection. / Oksuz, Ilkay; Ruijsink, Bram; Puyol-Antón, Esther et al.
Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings. ed. / Julia A. Schnabel; Christos Davatzikos; Carlos Alberola-López; Gabor Fichtinger; Alejandro F. Frangi. Springer Verlag, 2018. p. 250-258 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11070 LNCS).

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

TY - GEN

T1 - Deep learning using K-space based data augmentation for automated cardiac MR motion artefact detection

AU - Oksuz, Ilkay

AU - Ruijsink, Bram

AU - Puyol-Antón, Esther

AU - Bustin, Aurelien

AU - Cruz, Gastao

AU - Prieto, Claudia

AU - Rueckert, Daniel

AU - Schnabel, Julia A.

AU - King, Andrew P.

N1 - Publisher Copyright: © Springer Nature Switzerland AG 2018.

PY - 2018

Y1 - 2018

N2 - Quality assessment of medical images is essential for complete automation of image processing pipelines. For large population studies such as the UK Biobank, artefacts such as those caused by heart motion are problematic and manual identification is tedious and time-consuming. Therefore, there is an urgent need for automatic image quality assessment techniques. In this paper, we propose a method to automatically detect the presence of motion-related artefacts in cardiac magnetic resonance (CMR) images. As this is a highly imbalanced classification problem (due to the high number of good quality images compared to the low number of images with motion artefacts), we propose a novel k-space based training data augmentation approach in order to address this problem. Our method is based on 3D spatio-temporal Convolutional Neural Networks, and is able to detect 2D+time short axis images with motion artefacts in less than 1 ms. We test our algorithm on a subset of the UK Biobank dataset consisting of 3465 CMR images and achieve not only high accuracy in detection of motion artefacts, but also high precision and recall. We compare our approach to a range of state-of-the-art quality assessment methods.

AB - Quality assessment of medical images is essential for complete automation of image processing pipelines. For large population studies such as the UK Biobank, artefacts such as those caused by heart motion are problematic and manual identification is tedious and time-consuming. Therefore, there is an urgent need for automatic image quality assessment techniques. In this paper, we propose a method to automatically detect the presence of motion-related artefacts in cardiac magnetic resonance (CMR) images. As this is a highly imbalanced classification problem (due to the high number of good quality images compared to the low number of images with motion artefacts), we propose a novel k-space based training data augmentation approach in order to address this problem. Our method is based on 3D spatio-temporal Convolutional Neural Networks, and is able to detect 2D+time short axis images with motion artefacts in less than 1 ms. We test our algorithm on a subset of the UK Biobank dataset consisting of 3465 CMR images and achieve not only high accuracy in detection of motion artefacts, but also high precision and recall. We compare our approach to a range of state-of-the-art quality assessment methods.

KW - Cardiac MR

KW - Convolutional Neural Networks

KW - Image quality assessment

KW - Motion artefacts

KW - UK Biobank

UR - https://www.scopus.com/pages/publications/85054097617

U2 - 10.1007/978-3-030-00928-1_29

DO - 10.1007/978-3-030-00928-1_29

M3 - Conference contribution

AN - SCOPUS:85054097617

SN - 9783030009274

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

SP - 250

EP - 258

BT - Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings

A2 - Schnabel, Julia A.

A2 - Davatzikos, Christos

A2 - Alberola-López, Carlos

A2 - Fichtinger, Gabor

A2 - Frangi, Alejandro F.

PB - Springer Verlag

T2 - 21st International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2018

Y2 - 16 September 2018 through 20 September 2018

ER -

Oksuz I, Ruijsink B, Puyol-Antón E, Bustin A, Cruz G, Prieto C et al. Deep learning using K-space based data augmentation for automated cardiac MR motion artefact detection. In Schnabel JA, Davatzikos C, Alberola-López C, Fichtinger G, Frangi AF, editors, Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings. Springer Verlag. 2018. p. 250-258. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-00928-1_29

Deep learning using K-space based data augmentation for automated cardiac MR motion artefact detection

Abstract

Publication series

Conference

Bibliographical note

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this