Domain-adversarial neural networks to address the appearance variability of histopathology images

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/CongresprocedureConferentiebijdrageAcademicpeer review

20 Citaties (Scopus)
2 Downloads (Pure)

Uittreksel

Preparing and scanning histopathology slides consists of several steps, each with a multitude of parameters. The parameters can vary between pathology labs and within the same lab over time, resulting in significant variability of the tissue appearance that hampers the generalization of automatic image analysis methods. Typically, this is addressed with ad-hoc approaches such as staining normalization that aim to reduce the appearance variability. In this paper, we propose a systematic solution based on domain-adversarial neural networks. We hypothesize that removing the domain information from the model representation leads to better generalization. We tested our hypothesis for the problem of mitosis detection in breast cancer histopathology images and made a comparative analysis with two other approaches. We show that combining color augmentation with domain-adversarial training is a better alternative than standard approaches to improve the generalization of deep learning methods.

Originele taal-2Engels
TitelDeep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support - 3rd International Workshop, DLMIA 2017 and 7th International Workshop, ML-CDS 2017 Held in Conjunction with MICCAI 2017, Proceedings
UitgeverijSpringer
Pagina's83-91
Aantal pagina's9
ISBN van geprinte versie9783319675572
DOI's
StatusGepubliceerd - 2017
Evenement3rd International Workshop on Deep Learning in Medical Image Analysis, DLMIA 2017 and 7th International Workshop on Multimodal Learning for Clinical Decision Support, ML-CDS 2017 held in Conjunction with 20th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2017, 10-14 September Quebec, Canada - Quebec City, Canada
Duur: 10 sep 201714 sep 2017
http://www.miccai2017.org/

Publicatie series

NaamLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10553 LNCS
ISSN van geprinte versie0302-9743
ISSN van elektronische versie1611-3349

Congres

Congres3rd International Workshop on Deep Learning in Medical Image Analysis, DLMIA 2017 and 7th International Workshop on Multimodal Learning for Clinical Decision Support, ML-CDS 2017 held in Conjunction with 20th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2017, 10-14 September Quebec, Canada
Verkorte titelDLMIA2017
LandCanada
StadQuebec City
Periode10/09/1714/09/17
Internet adres

Vingerafdruk

Pathology
Image analysis
Neural Networks
Tissue
Color
Neural networks
Scanning
Augmentation
Breast Cancer
Comparative Analysis
Image Analysis
Normalization
Vary
Alternatives
Deep learning
Generalization
Model

Citeer dit

Lafarge, M. W., Pluim, J. P. W., Eppenhof, K. A. J., Moeskops, P., & Veta, M. (2017). Domain-adversarial neural networks to address the appearance variability of histopathology images. In Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support - 3rd International Workshop, DLMIA 2017 and 7th International Workshop, ML-CDS 2017 Held in Conjunction with MICCAI 2017, Proceedings (blz. 83-91). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10553 LNCS). Springer. https://doi.org/10.1007/978-3-319-67558-9_10
Lafarge, M.W. ; Pluim, J.P.W. ; Eppenhof, K.A.J. ; Moeskops, P. ; Veta, M. / Domain-adversarial neural networks to address the appearance variability of histopathology images. Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support - 3rd International Workshop, DLMIA 2017 and 7th International Workshop, ML-CDS 2017 Held in Conjunction with MICCAI 2017, Proceedings. Springer, 2017. blz. 83-91 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{7293afeac1d148018c1ef2e4b35f3597,
title = "Domain-adversarial neural networks to address the appearance variability of histopathology images",
abstract = "Preparing and scanning histopathology slides consists of several steps, each with a multitude of parameters. The parameters can vary between pathology labs and within the same lab over time, resulting in significant variability of the tissue appearance that hampers the generalization of automatic image analysis methods. Typically, this is addressed with ad-hoc approaches such as staining normalization that aim to reduce the appearance variability. In this paper, we propose a systematic solution based on domain-adversarial neural networks. We hypothesize that removing the domain information from the model representation leads to better generalization. We tested our hypothesis for the problem of mitosis detection in breast cancer histopathology images and made a comparative analysis with two other approaches. We show that combining color augmentation with domain-adversarial training is a better alternative than standard approaches to improve the generalization of deep learning methods.",
keywords = "Domain-adversarial training, Histopathology image analysis",
author = "M.W. Lafarge and J.P.W. Pluim and K.A.J. Eppenhof and P. Moeskops and M. Veta",
year = "2017",
doi = "10.1007/978-3-319-67558-9_10",
language = "English",
isbn = "9783319675572",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer",
pages = "83--91",
booktitle = "Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support - 3rd International Workshop, DLMIA 2017 and 7th International Workshop, ML-CDS 2017 Held in Conjunction with MICCAI 2017, Proceedings",
address = "Germany",

}

Lafarge, MW, Pluim, JPW, Eppenhof, KAJ, Moeskops, P & Veta, M 2017, Domain-adversarial neural networks to address the appearance variability of histopathology images. in Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support - 3rd International Workshop, DLMIA 2017 and 7th International Workshop, ML-CDS 2017 Held in Conjunction with MICCAI 2017, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10553 LNCS, Springer, blz. 83-91, Quebec City, Canada, 10/09/17. https://doi.org/10.1007/978-3-319-67558-9_10

Domain-adversarial neural networks to address the appearance variability of histopathology images. / Lafarge, M.W.; Pluim, J.P.W.; Eppenhof, K.A.J.; Moeskops, P.; Veta, M.

Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support - 3rd International Workshop, DLMIA 2017 and 7th International Workshop, ML-CDS 2017 Held in Conjunction with MICCAI 2017, Proceedings. Springer, 2017. blz. 83-91 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10553 LNCS).

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/CongresprocedureConferentiebijdrageAcademicpeer review

TY - GEN

T1 - Domain-adversarial neural networks to address the appearance variability of histopathology images

AU - Lafarge, M.W.

AU - Pluim, J.P.W.

AU - Eppenhof, K.A.J.

AU - Moeskops, P.

AU - Veta, M.

PY - 2017

Y1 - 2017

N2 - Preparing and scanning histopathology slides consists of several steps, each with a multitude of parameters. The parameters can vary between pathology labs and within the same lab over time, resulting in significant variability of the tissue appearance that hampers the generalization of automatic image analysis methods. Typically, this is addressed with ad-hoc approaches such as staining normalization that aim to reduce the appearance variability. In this paper, we propose a systematic solution based on domain-adversarial neural networks. We hypothesize that removing the domain information from the model representation leads to better generalization. We tested our hypothesis for the problem of mitosis detection in breast cancer histopathology images and made a comparative analysis with two other approaches. We show that combining color augmentation with domain-adversarial training is a better alternative than standard approaches to improve the generalization of deep learning methods.

AB - Preparing and scanning histopathology slides consists of several steps, each with a multitude of parameters. The parameters can vary between pathology labs and within the same lab over time, resulting in significant variability of the tissue appearance that hampers the generalization of automatic image analysis methods. Typically, this is addressed with ad-hoc approaches such as staining normalization that aim to reduce the appearance variability. In this paper, we propose a systematic solution based on domain-adversarial neural networks. We hypothesize that removing the domain information from the model representation leads to better generalization. We tested our hypothesis for the problem of mitosis detection in breast cancer histopathology images and made a comparative analysis with two other approaches. We show that combining color augmentation with domain-adversarial training is a better alternative than standard approaches to improve the generalization of deep learning methods.

KW - Domain-adversarial training

KW - Histopathology image analysis

UR - http://www.scopus.com/inward/record.url?scp=85029818275&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-67558-9_10

DO - 10.1007/978-3-319-67558-9_10

M3 - Conference contribution

AN - SCOPUS:85029818275

SN - 9783319675572

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

SP - 83

EP - 91

BT - Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support - 3rd International Workshop, DLMIA 2017 and 7th International Workshop, ML-CDS 2017 Held in Conjunction with MICCAI 2017, Proceedings

PB - Springer

ER -

Lafarge MW, Pluim JPW, Eppenhof KAJ, Moeskops P, Veta M. Domain-adversarial neural networks to address the appearance variability of histopathology images. In Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support - 3rd International Workshop, DLMIA 2017 and 7th International Workshop, ML-CDS 2017 Held in Conjunction with MICCAI 2017, Proceedings. Springer. 2017. blz. 83-91. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-67558-9_10