Learning an MR acquisition-invariant representation using Siamese neural networks

W.M. Kouw; M. Loog; L.W. Bartels; A.M. Mendrik

doi:10.1109/ISBI.2019.8759281

Learning an MR acquisition-invariant representation using Siamese neural networks

W.M. Kouw, M. Loog, L.W. Bartels, A.M. Mendrik

Pattern Recognition and Bioinformatics

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

4 Citations (Scopus)

Abstract

Generalization of voxelwise classifiers is hampered by differences between MRI-scanners, e.g. different acquisition protocols and field strengths. To address this limitation, we propose a Siamese neural network (MRAI-net) that extracts acquisition-invariant feature vectors. These can consequently be used by task-specific methods, such as voxelwise classifiers for tissue segmentation. MRAI-net is tested on both simulated and real patient data. Experiments show that MRAI-net outperforms voxelwise classifiers trained on the source or target scanner data when a small number of labeled samples is available.

Original language	English
Title of host publication	2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019)
Subtitle of host publication	Proceedings
Place of Publication	Danvers
Publisher	IEEE
Pages	364-367
Number of pages	4
ISBN (Electronic)	978-1-5386-3641-1
ISBN (Print)	978-1-5386-3642-8
DOIs	https://doi.org/10.1109/ISBI.2019.8759281
Publication status	Published - 2019
Event	IEEE International Symposium on Biomedical Imaging, ISBI 2019 - Venice, Italy Duration: 8 Apr 2019 → 11 Apr 2019 https://biomedicalimaging.org/2019/

Conference

Conference	IEEE International Symposium on Biomedical Imaging, ISBI 2019
Abbreviated title	ISBI'19
Country/Territory	Italy
City	Venice
Period	8/04/19 → 11/04/19
Internet address	https://biomedicalimaging.org/2019/

Keywords

MRI
Acquisition-variation
Representation-learning
Siamese-neural-network

Access to Document

10.1109/ISBI.2019.8759281

Cite this

@inproceedings{cbd6ed9e9df44894bb27fccfc4f96eb1,

title = "Learning an MR acquisition-invariant representation using Siamese neural networks",

abstract = "Generalization of voxelwise classifiers is hampered by differences between MRI-scanners, e.g. different acquisition protocols and field strengths. To address this limitation, we propose a Siamese neural network (MRAI-net) that extracts acquisition-invariant feature vectors. These can consequently be used by task-specific methods, such as voxelwise classifiers for tissue segmentation. MRAI-net is tested on both simulated and real patient data. Experiments show that MRAI-net outperforms voxelwise classifiers trained on the source or target scanner data when a small number of labeled samples is available.",

keywords = "MRI, Acquisition-variation, Representation-learning, Siamese-neural-network",

author = "W.M. Kouw and M. Loog and L.W. Bartels and A.M. Mendrik",

year = "2019",

doi = "10.1109/ISBI.2019.8759281",

language = "English",

isbn = "978-1-5386-3642-8 ",

pages = "364--367",

booktitle = "2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019)",

publisher = "IEEE",

address = "United States",

note = "IEEE International Symposium on Biomedical Imaging, ISBI 2019, ISBI'19 ; Conference date: 08-04-2019 Through 11-04-2019",

url = "https://biomedicalimaging.org/2019/",

}

Kouw, WM, Loog, M, Bartels, LW & Mendrik, AM 2019, Learning an MR acquisition-invariant representation using Siamese neural networks. in 2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019) : Proceedings. IEEE, Danvers, pp. 364-367, IEEE International Symposium on Biomedical Imaging, ISBI 2019, Venice, Italy, 8/04/19. https://doi.org/10.1109/ISBI.2019.8759281

Learning an MR acquisition-invariant representation using Siamese neural networks. / Kouw, W.M.; Loog, M.; Bartels, L.W. et al.
2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019) : Proceedings. Danvers: IEEE, 2019. p. 364-367.

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

TY - GEN

T1 - Learning an MR acquisition-invariant representation using Siamese neural networks

AU - Kouw, W.M.

AU - Loog, M.

AU - Bartels, L.W.

AU - Mendrik, A.M.

PY - 2019

Y1 - 2019

N2 - Generalization of voxelwise classifiers is hampered by differences between MRI-scanners, e.g. different acquisition protocols and field strengths. To address this limitation, we propose a Siamese neural network (MRAI-net) that extracts acquisition-invariant feature vectors. These can consequently be used by task-specific methods, such as voxelwise classifiers for tissue segmentation. MRAI-net is tested on both simulated and real patient data. Experiments show that MRAI-net outperforms voxelwise classifiers trained on the source or target scanner data when a small number of labeled samples is available.

AB - Generalization of voxelwise classifiers is hampered by differences between MRI-scanners, e.g. different acquisition protocols and field strengths. To address this limitation, we propose a Siamese neural network (MRAI-net) that extracts acquisition-invariant feature vectors. These can consequently be used by task-specific methods, such as voxelwise classifiers for tissue segmentation. MRAI-net is tested on both simulated and real patient data. Experiments show that MRAI-net outperforms voxelwise classifiers trained on the source or target scanner data when a small number of labeled samples is available.

KW - MRI

KW - Acquisition-variation

KW - Representation-learning

KW - Siamese-neural-network

U2 - 10.1109/ISBI.2019.8759281

DO - 10.1109/ISBI.2019.8759281

M3 - Conference contribution

SN - 978-1-5386-3642-8

SP - 364

EP - 367

BT - 2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019)

PB - IEEE

CY - Danvers

T2 - IEEE International Symposium on Biomedical Imaging, ISBI 2019

Y2 - 8 April 2019 through 11 April 2019

ER -

Learning an MR acquisition-invariant representation using Siamese neural networks

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this