Robust Segmentation of Overlapping Cells in Cervical Cytology Using Light Convolution Neural Network

Shusong Xu; Chen Sang; Yulan Jin; Tao Wan

doi:10.1007/978-3-030-04239-4_35

Robust Segmentation of Overlapping Cells in Cervical Cytology Using Light Convolution Neural Network

Shusong Xu
, Chen Sang
, Yulan Jin
, Tao Wan^*

^*Corresponding author for this work

School of Biological Science and Medical Engineering

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

3 Scopus citations

Abstract

Automated segmentation of cells in cervical cytology images poses a great challenge due to the presence of fuzzy and overlapping cells, noisy background, and poor cytoplasmic contrast. We present an improved method for segmenting nuclei and cytoplasm from a cluster of cervical cells using convolutional network and fast multi-cell labeling. A light convolutional neural network (CNN) model is employed to generate nuclei candidates, which can serve as accurate initializations for the subsequent level set segmentation and provide a priori knowledge for the cytoplasm segmentation. A fast multi-cell labeling method based on the superpixel map is devised to roughly segment clumped and inhomogeneous cytoplasm before applying a cell boundary refinement approach. A shape constraint in conjunction with boundary and region information drive a level set formulation to perform a robust cell segmentation. The qualitative and quantitative evaluations demonstrated that the presented cellular segmentation method is effective and efficient.

Original language	English
Title of host publication	Neural Information Processing - 25th International Conference, ICONIP 2018, Proceedings
Editors	Seiichi Ozawa, Andrew Chi Sing Leung, Long Cheng
Publisher	Springer Verlag
Pages	387-397
Number of pages	11
ISBN (Print)	9783030042387
DOIs	https://doi.org/10.1007/978-3-030-04239-4_35
State	Published - 2018
Event	25th International Conference on Neural Information Processing, ICONIP 2018 - Siem Reap, Cambodia Duration: 13 Dec 2018 → 16 Dec 2018

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11307 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	25th International Conference on Neural Information Processing, ICONIP 2018
Country/Territory	Cambodia
City	Siem Reap
Period	13/12/18 → 16/12/18

Keywords

Cervical cytology
Convolutional Neural Network
Cytoplasm segmentation
Multi-cell labeling
Nuclei detection

Access to Document

10.1007/978-3-030-04239-4_35

Cite this

Xu, S., Sang, C., Jin, Y., & Wan, T. (2018). Robust Segmentation of Overlapping Cells in Cervical Cytology Using Light Convolution Neural Network. In S. Ozawa, A. C. S. Leung, & L. Cheng (Eds.), Neural Information Processing - 25th International Conference, ICONIP 2018, Proceedings (pp. 387-397). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11307 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-04239-4_35

Xu, Shusong ; Sang, Chen ; Jin, Yulan et al. / Robust Segmentation of Overlapping Cells in Cervical Cytology Using Light Convolution Neural Network. Neural Information Processing - 25th International Conference, ICONIP 2018, Proceedings. editor / Seiichi Ozawa ; Andrew Chi Sing Leung ; Long Cheng. Springer Verlag, 2018. pp. 387-397 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{f1bd1e4aad0740e9949bc46c2586da82,

title = "Robust Segmentation of Overlapping Cells in Cervical Cytology Using Light Convolution Neural Network",

abstract = "Automated segmentation of cells in cervical cytology images poses a great challenge due to the presence of fuzzy and overlapping cells, noisy background, and poor cytoplasmic contrast. We present an improved method for segmenting nuclei and cytoplasm from a cluster of cervical cells using convolutional network and fast multi-cell labeling. A light convolutional neural network (CNN) model is employed to generate nuclei candidates, which can serve as accurate initializations for the subsequent level set segmentation and provide a priori knowledge for the cytoplasm segmentation. A fast multi-cell labeling method based on the superpixel map is devised to roughly segment clumped and inhomogeneous cytoplasm before applying a cell boundary refinement approach. A shape constraint in conjunction with boundary and region information drive a level set formulation to perform a robust cell segmentation. The qualitative and quantitative evaluations demonstrated that the presented cellular segmentation method is effective and efficient.",

keywords = "Cervical cytology, Convolutional Neural Network, Cytoplasm segmentation, Multi-cell labeling, Nuclei detection",

author = "Shusong Xu and Chen Sang and Yulan Jin and Tao Wan",

note = "Publisher Copyright: {\textcopyright} 2018, Springer Nature Switzerland AG.; 25th International Conference on Neural Information Processing, ICONIP 2018 ; Conference date: 13-12-2018 Through 16-12-2018",

year = "2018",

doi = "10.1007/978-3-030-04239-4\_35",

language = "英语",

isbn = "9783030042387",

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

publisher = "Springer Verlag",

pages = "387--397",

editor = "Seiichi Ozawa and Leung, \{Andrew Chi Sing\} and Long Cheng",

booktitle = "Neural Information Processing - 25th International Conference, ICONIP 2018, Proceedings",

address = "德国",

}

Xu, S, Sang, C, Jin, Y & Wan, T 2018, Robust Segmentation of Overlapping Cells in Cervical Cytology Using Light Convolution Neural Network. in S Ozawa, ACS Leung & L Cheng (eds), Neural Information Processing - 25th International Conference, ICONIP 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11307 LNCS, Springer Verlag, pp. 387-397, 25th International Conference on Neural Information Processing, ICONIP 2018, Siem Reap, Cambodia, 13/12/18. https://doi.org/10.1007/978-3-030-04239-4_35

Robust Segmentation of Overlapping Cells in Cervical Cytology Using Light Convolution Neural Network. / Xu, Shusong; Sang, Chen; Jin, Yulan et al.
Neural Information Processing - 25th International Conference, ICONIP 2018, Proceedings. ed. / Seiichi Ozawa; Andrew Chi Sing Leung; Long Cheng. Springer Verlag, 2018. p. 387-397 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11307 LNCS).

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

TY - GEN

T1 - Robust Segmentation of Overlapping Cells in Cervical Cytology Using Light Convolution Neural Network

AU - Xu, Shusong

AU - Sang, Chen

AU - Jin, Yulan

AU - Wan, Tao

PY - 2018

Y1 - 2018

N2 - Automated segmentation of cells in cervical cytology images poses a great challenge due to the presence of fuzzy and overlapping cells, noisy background, and poor cytoplasmic contrast. We present an improved method for segmenting nuclei and cytoplasm from a cluster of cervical cells using convolutional network and fast multi-cell labeling. A light convolutional neural network (CNN) model is employed to generate nuclei candidates, which can serve as accurate initializations for the subsequent level set segmentation and provide a priori knowledge for the cytoplasm segmentation. A fast multi-cell labeling method based on the superpixel map is devised to roughly segment clumped and inhomogeneous cytoplasm before applying a cell boundary refinement approach. A shape constraint in conjunction with boundary and region information drive a level set formulation to perform a robust cell segmentation. The qualitative and quantitative evaluations demonstrated that the presented cellular segmentation method is effective and efficient.

AB - Automated segmentation of cells in cervical cytology images poses a great challenge due to the presence of fuzzy and overlapping cells, noisy background, and poor cytoplasmic contrast. We present an improved method for segmenting nuclei and cytoplasm from a cluster of cervical cells using convolutional network and fast multi-cell labeling. A light convolutional neural network (CNN) model is employed to generate nuclei candidates, which can serve as accurate initializations for the subsequent level set segmentation and provide a priori knowledge for the cytoplasm segmentation. A fast multi-cell labeling method based on the superpixel map is devised to roughly segment clumped and inhomogeneous cytoplasm before applying a cell boundary refinement approach. A shape constraint in conjunction with boundary and region information drive a level set formulation to perform a robust cell segmentation. The qualitative and quantitative evaluations demonstrated that the presented cellular segmentation method is effective and efficient.

KW - Cervical cytology

KW - Convolutional Neural Network

KW - Cytoplasm segmentation

KW - Multi-cell labeling

KW - Nuclei detection

UR - https://www.scopus.com/pages/publications/85059005349

U2 - 10.1007/978-3-030-04239-4_35

DO - 10.1007/978-3-030-04239-4_35

M3 - 会议稿件

AN - SCOPUS:85059005349

SN - 9783030042387

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

SP - 387

EP - 397

BT - Neural Information Processing - 25th International Conference, ICONIP 2018, Proceedings

A2 - Ozawa, Seiichi

A2 - Leung, Andrew Chi Sing

A2 - Cheng, Long

PB - Springer Verlag

T2 - 25th International Conference on Neural Information Processing, ICONIP 2018

Y2 - 13 December 2018 through 16 December 2018

ER -

Xu S, Sang C, Jin Y, Wan T. Robust Segmentation of Overlapping Cells in Cervical Cytology Using Light Convolution Neural Network. In Ozawa S, Leung ACS, Cheng L, editors, Neural Information Processing - 25th International Conference, ICONIP 2018, Proceedings. Springer Verlag. 2018. p. 387-397. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-04239-4_35

Robust Segmentation of Overlapping Cells in Cervical Cytology Using Light Convolution Neural Network

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this