Stain Colour Normalisation to Improve Mitosis Detection on Breast Histology Images

Azam Hamidinekoo; Reyer Zwiggelaar

doi:10.1007/978-3-319-67558-9_25

Stain Colour Normalisation to Improve Mitosis Detection on Breast Histology Images

Azam Hamidinekoo
, Reyer Zwiggelaar

Department of Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Conference Proceeding (ISBN)

5 Citations (Scopus)

Abstract

The mitosis count, one of the main components considered for grading breast cancer on histology images, is used to assess tumour proliferation. On breast histology sections, different coloured stains are used to highlight existing cellular components. To keep the wealth of information in the stain colour representation and decrease the sensitivity of detection and classification models to stain variations, the RGB Histogram Specification method is used as a preprocessing step in the training of a modified deep convolutional neural network. Different models are trained on raw and stain normalised images using different databases. Evaluation results show more stable detection performance for various imaging conditions. Combining different data sources and employing stain colour normalisation and transfer learning, a network is trained that can be used for the general mitosis detection task and dealing with staining and scanner variations.

Original language	English
Title of host publication	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
Subtitle of host publication	Third International Workshop, DLMIA 2017, and 7th International Workshop, ML-CDS 2017, Held in Conjunction with MICCAI 2017, Québec City, QC, Canada, September 14, Proceedings
Editors	M. Jorge Cardoso, Tal Arbel, Gustavo Carneiro, João Manuel R.S. Tavares, Mehdi Moradi, Andrew Bradley, Hayit Greenspan, João Paulo Papa, Arnant Madabhushi, Jacinto C. Nascimento, Jaime S. Cardoso, Vasileios Belagiannis, Zhi Lu
Publisher	Springer Nature
Pages	213-221
Number of pages	9
ISBN (Electronic)	978-3-319-67558-9
ISBN (Print)	978-3-319-67557-2
DOIs	https://doi.org/10.1007/978-3-319-67558-9_25
Publication status	Published - 09 Sept 2017
Event	Third International Workshop, DLMIA 2017, and 7th International Workshop, ML-CDS 2017, Held in Conjunction with MICCAI 2017 - Québec City, Canada Duration: 14 Sept 2017 → …

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	10553 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	Third International Workshop, DLMIA 2017, and 7th International Workshop, ML-CDS 2017, Held in Conjunction with MICCAI 2017
Country/Territory	Canada
City	Québec City
Period	14 Sept 2017 → …

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 3 Good Health and Well-being

Access to Document

10.1007/978-3-319-67558-9_25

Permanent link

Persistent link

Cite this

Hamidinekoo, A., & Zwiggelaar, R. (2017). Stain Colour Normalisation to Improve Mitosis Detection on Breast Histology Images. In M. J. Cardoso, T. Arbel, G. Carneiro, J. M. R. S. Tavares, M. Moradi, A. Bradley, H. Greenspan, J. P. Papa, A. Madabhushi, J. C. Nascimento, J. S. Cardoso, V. Belagiannis, & Z. Lu (Eds.), 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: Third International Workshop, DLMIA 2017, and 7th International Workshop, ML-CDS 2017, Held in Conjunction with MICCAI 2017, Québec City, QC, Canada, September 14, Proceedings (pp. 213-221). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10553 LNCS). Springer Nature. https://doi.org/10.1007/978-3-319-67558-9_25

Hamidinekoo, Azam ; Zwiggelaar, Reyer. / Stain Colour Normalisation to Improve Mitosis Detection on Breast Histology 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: Third International Workshop, DLMIA 2017, and 7th International Workshop, ML-CDS 2017, Held in Conjunction with MICCAI 2017, Québec City, QC, Canada, September 14, Proceedings. editor / M. Jorge Cardoso ; Tal Arbel ; Gustavo Carneiro ; João Manuel R.S. Tavares ; Mehdi Moradi ; Andrew Bradley ; Hayit Greenspan ; João Paulo Papa ; Arnant Madabhushi ; Jacinto C. Nascimento ; Jaime S. Cardoso ; Vasileios Belagiannis ; Zhi Lu. Springer Nature, 2017. pp. 213-221 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{0aaa0eaf502a48d3ad907b300e0ef3e2,

title = "Stain Colour Normalisation to Improve Mitosis Detection on Breast Histology Images",

abstract = "The mitosis count, one of the main components considered for grading breast cancer on histology images, is used to assess tumour proliferation. On breast histology sections, different coloured stains are used to highlight existing cellular components. To keep the wealth of information in the stain colour representation and decrease the sensitivity of detection and classification models to stain variations, the RGB Histogram Specification method is used as a preprocessing step in the training of a modified deep convolutional neural network. Different models are trained on raw and stain normalised images using different databases. Evaluation results show more stable detection performance for various imaging conditions. Combining different data sources and employing stain colour normalisation and transfer learning, a network is trained that can be used for the general mitosis detection task and dealing with staining and scanner variations.",

author = "Azam Hamidinekoo and Reyer Zwiggelaar",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing AG 2017.; Third International Workshop, DLMIA 2017, and 7th International Workshop, ML-CDS 2017, Held in Conjunction with MICCAI 2017 ; Conference date: 14-09-2017",

year = "2017",

month = sep,

day = "9",

doi = "10.1007/978-3-319-67558-9\_25",

language = "English",

isbn = "978-3-319-67557-2",

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

publisher = "Springer Nature",

pages = "213--221",

editor = "Cardoso, \{M. Jorge\} and Tal Arbel and Gustavo Carneiro and Tavares, \{Jo{\~a}o Manuel R.S.\} and Mehdi Moradi and Andrew Bradley and Hayit Greenspan and Papa, \{Jo{\~a}o Paulo\} and Arnant Madabhushi and Nascimento, \{Jacinto C.\} and Cardoso, \{Jaime S.\} and Vasileios Belagiannis and Zhi Lu",

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 = "Switzerland",

}

Hamidinekoo, A & Zwiggelaar, R 2017, Stain Colour Normalisation to Improve Mitosis Detection on Breast Histology Images. in MJ Cardoso, T Arbel, G Carneiro, JMRS Tavares, M Moradi, A Bradley, H Greenspan, JP Papa, A Madabhushi, JC Nascimento, JS Cardoso, V Belagiannis & Z Lu (eds), 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: Third International Workshop, DLMIA 2017, and 7th International Workshop, ML-CDS 2017, Held in Conjunction with MICCAI 2017, Québec City, QC, Canada, September 14, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10553 LNCS, Springer Nature, pp. 213-221, Third International Workshop, DLMIA 2017, and 7th International Workshop, ML-CDS 2017, Held in Conjunction with MICCAI 2017, Québec City, Quebec, Canada, 14 Sept 2017. https://doi.org/10.1007/978-3-319-67558-9_25

Stain Colour Normalisation to Improve Mitosis Detection on Breast Histology Images. / Hamidinekoo, Azam; Zwiggelaar, Reyer.
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: Third International Workshop, DLMIA 2017, and 7th International Workshop, ML-CDS 2017, Held in Conjunction with MICCAI 2017, Québec City, QC, Canada, September 14, Proceedings. ed. / M. Jorge Cardoso; Tal Arbel; Gustavo Carneiro; João Manuel R.S. Tavares; Mehdi Moradi; Andrew Bradley; Hayit Greenspan; João Paulo Papa; Arnant Madabhushi; Jacinto C. Nascimento; Jaime S. Cardoso; Vasileios Belagiannis; Zhi Lu. Springer Nature, 2017. p. 213-221 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10553 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference Proceeding (ISBN)

TY - GEN

T1 - Stain Colour Normalisation to Improve Mitosis Detection on Breast Histology Images

AU - Hamidinekoo, Azam

AU - Zwiggelaar, Reyer

N1 - Publisher Copyright: © Springer International Publishing AG 2017.

PY - 2017/9/9

Y1 - 2017/9/9

N2 - The mitosis count, one of the main components considered for grading breast cancer on histology images, is used to assess tumour proliferation. On breast histology sections, different coloured stains are used to highlight existing cellular components. To keep the wealth of information in the stain colour representation and decrease the sensitivity of detection and classification models to stain variations, the RGB Histogram Specification method is used as a preprocessing step in the training of a modified deep convolutional neural network. Different models are trained on raw and stain normalised images using different databases. Evaluation results show more stable detection performance for various imaging conditions. Combining different data sources and employing stain colour normalisation and transfer learning, a network is trained that can be used for the general mitosis detection task and dealing with staining and scanner variations.

AB - The mitosis count, one of the main components considered for grading breast cancer on histology images, is used to assess tumour proliferation. On breast histology sections, different coloured stains are used to highlight existing cellular components. To keep the wealth of information in the stain colour representation and decrease the sensitivity of detection and classification models to stain variations, the RGB Histogram Specification method is used as a preprocessing step in the training of a modified deep convolutional neural network. Different models are trained on raw and stain normalised images using different databases. Evaluation results show more stable detection performance for various imaging conditions. Combining different data sources and employing stain colour normalisation and transfer learning, a network is trained that can be used for the general mitosis detection task and dealing with staining and scanner variations.

UR - https://www.scopus.com/pages/publications/85029819818

U2 - 10.1007/978-3-319-67558-9_25

DO - 10.1007/978-3-319-67558-9_25

M3 - Conference Proceeding (ISBN)

SN - 978-3-319-67557-2

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

SP - 213

EP - 221

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

A2 - Cardoso, M. Jorge

A2 - Arbel, Tal

A2 - Carneiro, Gustavo

A2 - Tavares, João Manuel R.S.

A2 - Moradi, Mehdi

A2 - Bradley, Andrew

A2 - Greenspan, Hayit

A2 - Papa, João Paulo

A2 - Madabhushi, Arnant

A2 - Nascimento, Jacinto C.

A2 - Cardoso, Jaime S.

A2 - Belagiannis, Vasileios

A2 - Lu, Zhi

PB - Springer Nature

T2 - Third International Workshop, DLMIA 2017, and 7th International Workshop, ML-CDS 2017, Held in Conjunction with MICCAI 2017

Y2 - 14 September 2017

ER -

Hamidinekoo A, Zwiggelaar R. Stain Colour Normalisation to Improve Mitosis Detection on Breast Histology Images. In Cardoso MJ, Arbel T, Carneiro G, Tavares JMRS, Moradi M, Bradley A, Greenspan H, Papa JP, Madabhushi A, Nascimento JC, Cardoso JS, Belagiannis V, Lu Z, editors, 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: Third International Workshop, DLMIA 2017, and 7th International Workshop, ML-CDS 2017, Held in Conjunction with MICCAI 2017, Québec City, QC, Canada, September 14, Proceedings. Springer Nature. 2017. p. 213-221. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). Epub 2017 Sept 7. doi: 10.1007/978-3-319-67558-9_25

Stain Colour Normalisation to Improve Mitosis Detection on Breast Histology Images

Abstract

Publication series

Conference

UN SDGs

Access to Document

Permanent link

Other files and links

Fingerprint

Cite this