Unsupervised Deep Learning for Stain Separation and Artifact Detection in Histopathology Images

Andrew Moyes; Kun Zhang; Ming Ji; Huiyu Zhou; Danny Crookes

doi:10.1007/978-3-030-52791-4_18

Unsupervised Deep Learning for Stain Separation and Artifact Detection in Histopathology Images

Andrew Moyes, Kun Zhang^*, Ming Ji, Huiyu Zhou, Danny Crookes

^*Corresponding author for this work

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

3 Citations (Scopus)

Abstract

Stain separation is an important pre-processing technique used to aid automated analysis of histopathology images. In this paper, we propose a novel, unsupervised deep learning method for stain separation (Hematoxylin and Eosin). This approach is inspired by Non-Negative Matrix Factorisation (NMF) and decomposes an input image into a stain colour matrix and a stain concentration matrix. In contrast to existing approaches, our method predicts stain colour matrices at the pixel level rather than the image level, thus enabling implicit modelling of tissue-dependant interactions between stains. We demonstrate an 8.81% reduction in mean-squared error on a stain separation task measuring the similarity between predicted and actual hematoxylin images from a publicly available dataset of digitised tissue images. We also present a novel approach to artifact detection in histological images based on a constrained generative adversarial network which we demonstrate is able to detect a variety of artifact types without the use of labels.

Original language	English
Title of host publication	Medical Image Understanding and Analysis - 24th Annual Conference, MIUA 2020, Proceedings
Editors	Bartlomiej W. Papiez, Ana I.L. Namburete, Mohammad Yaqub, J. Alison Noble, Mohammad Yaqub
Publisher	Springer
Pages	221-234
Number of pages	14
ISBN (Print)	9783030527907
DOIs	https://doi.org/10.1007/978-3-030-52791-4_18
Publication status	Published - 08 Jul 2020
Event	24th Annual Conference on Medical Image Understanding and Analysis, MIUA 2020 - Oxford, United Kingdom Duration: 15 Jul 2020 → 17 Jul 2020

Publication series

Name	Communications in Computer and Information Science
Volume	1248 CCIS
ISSN (Print)	1865-0929
ISSN (Electronic)	1865-0937

Conference

Conference	24th Annual Conference on Medical Image Understanding and Analysis, MIUA 2020
Country/Territory	United Kingdom
City	Oxford
Period	15/07/2020 → 17/07/2020

Bibliographical note

Funding Information:
This work was financially supported by Invest NI, the Natural Science Foundation of Jiangsu Province, China under Grant No. BK20170443 and the Natural Science Foundation of the Higher Education Institutions of Jiangsu Province, China under Grant No. 17KJB520030.

Publisher Copyright:
© 2020, Springer Nature Switzerland AG.

Copyright:
Copyright 2020 Elsevier B.V., All rights reserved.

Keywords

Artifact detection
Stain separation
Unsupervised

ASJC Scopus subject areas

General Computer Science
General Mathematics

Access to Document

10.1007/978-3-030-52791-4_18Licence: Unspecified

Best Paper
Moyes, Andrew (Recipient), 2020
Prize: Prize (including medals and awards)

Deep learning for processing histopathology images
Moyes, A. (Author), Ji, M. (Supervisor) & Gault, R. (Supervisor), Jul 2021
Student thesis: Doctoral Thesis › Doctor of Philosophy

File

3 Citations
1 Article

Multi-channel auto-encoders for learning domain invariant representations enabling superior classification of histopathology images
Moyes, A., Gault, R., Zhang, K., Ming, J., Crookes, D. & Wang, J., Jan 2023, In: Medical Image Analysis. 83, 10 p., 102640.
Research output: Contribution to journal › Article › peer-review

Open Access
File
15 Citations (Scopus)

136 Downloads (Pure)

Cite this

Moyes, A., Zhang, K., Ji, M., Zhou, H., & Crookes, D. (2020). Unsupervised Deep Learning for Stain Separation and Artifact Detection in Histopathology Images. In B. W. Papiez, A. I. L. Namburete, M. Yaqub, J. A. Noble, & M. Yaqub (Eds.), Medical Image Understanding and Analysis - 24th Annual Conference, MIUA 2020, Proceedings (pp. 221-234). (Communications in Computer and Information Science; Vol. 1248 CCIS). Springer. https://doi.org/10.1007/978-3-030-52791-4_18

Moyes, Andrew ; Zhang, Kun ; Ji, Ming et al. / Unsupervised Deep Learning for Stain Separation and Artifact Detection in Histopathology Images. Medical Image Understanding and Analysis - 24th Annual Conference, MIUA 2020, Proceedings. editor / Bartlomiej W. Papiez ; Ana I.L. Namburete ; Mohammad Yaqub ; J. Alison Noble ; Mohammad Yaqub. Springer, 2020. pp. 221-234 (Communications in Computer and Information Science).

@inproceedings{6dea73253c914d0fa5a02f1b906df713,

title = "Unsupervised Deep Learning for Stain Separation and Artifact Detection in Histopathology Images",

abstract = "Stain separation is an important pre-processing technique used to aid automated analysis of histopathology images. In this paper, we propose a novel, unsupervised deep learning method for stain separation (Hematoxylin and Eosin). This approach is inspired by Non-Negative Matrix Factorisation (NMF) and decomposes an input image into a stain colour matrix and a stain concentration matrix. In contrast to existing approaches, our method predicts stain colour matrices at the pixel level rather than the image level, thus enabling implicit modelling of tissue-dependant interactions between stains. We demonstrate an 8.81% reduction in mean-squared error on a stain separation task measuring the similarity between predicted and actual hematoxylin images from a publicly available dataset of digitised tissue images. We also present a novel approach to artifact detection in histological images based on a constrained generative adversarial network which we demonstrate is able to detect a variety of artifact types without the use of labels.",

keywords = "Artifact detection, Stain separation, Unsupervised",

author = "Andrew Moyes and Kun Zhang and Ming Ji and Huiyu Zhou and Danny Crookes",

note = "Funding Information: This work was financially supported by Invest NI, the Natural Science Foundation of Jiangsu Province, China under Grant No. BK20170443 and the Natural Science Foundation of the Higher Education Institutions of Jiangsu Province, China under Grant No. 17KJB520030. Publisher Copyright: {\textcopyright} 2020, Springer Nature Switzerland AG. Copyright: Copyright 2020 Elsevier B.V., All rights reserved.; 24th Annual Conference on Medical Image Understanding and Analysis, MIUA 2020 ; Conference date: 15-07-2020 Through 17-07-2020",

year = "2020",

month = jul,

day = "8",

doi = "10.1007/978-3-030-52791-4_18",

language = "English",

isbn = "9783030527907",

series = "Communications in Computer and Information Science",

publisher = "Springer",

pages = "221--234",

editor = "Papiez, {Bartlomiej W.} and Namburete, {Ana I.L.} and Mohammad Yaqub and Noble, {J. Alison} and Mohammad Yaqub",

booktitle = "Medical Image Understanding and Analysis - 24th Annual Conference, MIUA 2020, Proceedings",

}

Moyes, A, Zhang, K, Ji, M, Zhou, H & Crookes, D 2020, Unsupervised Deep Learning for Stain Separation and Artifact Detection in Histopathology Images. in BW Papiez, AIL Namburete, M Yaqub, JA Noble & M Yaqub (eds), Medical Image Understanding and Analysis - 24th Annual Conference, MIUA 2020, Proceedings. Communications in Computer and Information Science, vol. 1248 CCIS, Springer, pp. 221-234, 24th Annual Conference on Medical Image Understanding and Analysis, MIUA 2020, Oxford, United Kingdom, 15/07/2020. https://doi.org/10.1007/978-3-030-52791-4_18

Unsupervised Deep Learning for Stain Separation and Artifact Detection in Histopathology Images. / Moyes, Andrew; Zhang, Kun; Ji, Ming et al.
Medical Image Understanding and Analysis - 24th Annual Conference, MIUA 2020, Proceedings. ed. / Bartlomiej W. Papiez; Ana I.L. Namburete; Mohammad Yaqub; J. Alison Noble; Mohammad Yaqub. Springer, 2020. p. 221-234 (Communications in Computer and Information Science; Vol. 1248 CCIS).

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

TY - GEN

T1 - Unsupervised Deep Learning for Stain Separation and Artifact Detection in Histopathology Images

AU - Moyes, Andrew

AU - Zhang, Kun

AU - Ji, Ming

AU - Zhou, Huiyu

AU - Crookes, Danny

N1 - Funding Information: This work was financially supported by Invest NI, the Natural Science Foundation of Jiangsu Province, China under Grant No. BK20170443 and the Natural Science Foundation of the Higher Education Institutions of Jiangsu Province, China under Grant No. 17KJB520030. Publisher Copyright: © 2020, Springer Nature Switzerland AG. Copyright: Copyright 2020 Elsevier B.V., All rights reserved.

PY - 2020/7/8

Y1 - 2020/7/8

N2 - Stain separation is an important pre-processing technique used to aid automated analysis of histopathology images. In this paper, we propose a novel, unsupervised deep learning method for stain separation (Hematoxylin and Eosin). This approach is inspired by Non-Negative Matrix Factorisation (NMF) and decomposes an input image into a stain colour matrix and a stain concentration matrix. In contrast to existing approaches, our method predicts stain colour matrices at the pixel level rather than the image level, thus enabling implicit modelling of tissue-dependant interactions between stains. We demonstrate an 8.81% reduction in mean-squared error on a stain separation task measuring the similarity between predicted and actual hematoxylin images from a publicly available dataset of digitised tissue images. We also present a novel approach to artifact detection in histological images based on a constrained generative adversarial network which we demonstrate is able to detect a variety of artifact types without the use of labels.

AB - Stain separation is an important pre-processing technique used to aid automated analysis of histopathology images. In this paper, we propose a novel, unsupervised deep learning method for stain separation (Hematoxylin and Eosin). This approach is inspired by Non-Negative Matrix Factorisation (NMF) and decomposes an input image into a stain colour matrix and a stain concentration matrix. In contrast to existing approaches, our method predicts stain colour matrices at the pixel level rather than the image level, thus enabling implicit modelling of tissue-dependant interactions between stains. We demonstrate an 8.81% reduction in mean-squared error on a stain separation task measuring the similarity between predicted and actual hematoxylin images from a publicly available dataset of digitised tissue images. We also present a novel approach to artifact detection in histological images based on a constrained generative adversarial network which we demonstrate is able to detect a variety of artifact types without the use of labels.

KW - Artifact detection

KW - Stain separation

KW - Unsupervised

U2 - 10.1007/978-3-030-52791-4_18

DO - 10.1007/978-3-030-52791-4_18

M3 - Conference contribution

AN - SCOPUS:85088555086

SN - 9783030527907

T3 - Communications in Computer and Information Science

SP - 221

EP - 234

BT - Medical Image Understanding and Analysis - 24th Annual Conference, MIUA 2020, Proceedings

A2 - Papiez, Bartlomiej W.

A2 - Namburete, Ana I.L.

A2 - Yaqub, Mohammad

A2 - Noble, J. Alison

A2 - Yaqub, Mohammad

PB - Springer

T2 - 24th Annual Conference on Medical Image Understanding and Analysis, MIUA 2020

Y2 - 15 July 2020 through 17 July 2020

ER -

Moyes A, Zhang K, Ji M, Zhou H, Crookes D. Unsupervised Deep Learning for Stain Separation and Artifact Detection in Histopathology Images. In Papiez BW, Namburete AIL, Yaqub M, Noble JA, Yaqub M, editors, Medical Image Understanding and Analysis - 24th Annual Conference, MIUA 2020, Proceedings. Springer. 2020. p. 221-234. (Communications in Computer and Information Science). doi: 10.1007/978-3-030-52791-4_18

Unsupervised Deep Learning for Stain Separation and Artifact Detection in Histopathology Images

Abstract

Publication series

Conference

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Fingerprint

Prizes

Best Paper

Student theses

Deep learning for processing histopathology images

Research output

Multi-channel auto-encoders for learning domain invariant representations enabling superior classification of histopathology images

Cite this