Hard-Boundary Attention Network for Nuclei Instance Segmentation

Yalu Cheng; Pengchong Qiao; Hongliang He; Guoli Song; Jie Chen

doi:10.1145/3469877.3490602

Hard-Boundary Attention Network for Nuclei Instance Segmentation

Yalu Cheng, Pengchong Qiao, Hongliang He, Guoli Song, Jie Chen^*

^*Corresponding author for this work

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

3 Citations (Scopus)

Abstract

Image segmentation plays an important role in medical image analysis, and accurate segmentation of nuclei is especially crucial to clinical diagnosis. However, existing methods fail to segment dense nuclei due to the hard-boundary which has similar texture to nuclear inside. To this end, we propose a Hard-Boundary Attention Network (HBANet) for nuclei instance segmentation. Specifically, we propose a Background Weaken Module (BWM) to weaken the attention of our model to the nucleus background by integrating low-level features into high-level features. To improve the robustness of the model to the hard-boundary of nuclei, we further design a Gradient-based boundary adaptive Strategy (GS) which generates boundary-weakened data for model training in an adversarial manner. We conduct extensive experiments on MoNuSeg and CPM-17 datasets, and experimental results show that our HBANet outperforms the state-of-the-art methods.

Original language	English
Title of host publication	Proceedings of the 3rd ACM International Conference on Multimedia in Asia, MMAsia 2021
Publisher	Association for Computing Machinery
ISBN (Electronic)	9781450386074
DOIs	https://doi.org/10.1145/3469877.3490602
Publication status	Published - 1 Dec 2021
Externally published	Yes
Event	3rd ACM International Conference on Multimedia in Asia, MMAsia 2021 - Virtual, Online, Australia Duration: 1 Dec 2021 → 3 Dec 2021

Publication series

Name	ACM International Conference Proceeding Series

Conference

Conference	3rd ACM International Conference on Multimedia in Asia, MMAsia 2021
Country/Territory	Australia
City	Virtual, Online
Period	1/12/21 → 3/12/21

Keywords

Adaptive strategy
Deep learning
Nuclei instance segmentation

Access to Document

10.1145/3469877.3490602

Cite this

@inproceedings{526297f2461649038fa4a2fd84d7a4b8,

title = "Hard-Boundary Attention Network for Nuclei Instance Segmentation",

abstract = "Image segmentation plays an important role in medical image analysis, and accurate segmentation of nuclei is especially crucial to clinical diagnosis. However, existing methods fail to segment dense nuclei due to the hard-boundary which has similar texture to nuclear inside. To this end, we propose a Hard-Boundary Attention Network (HBANet) for nuclei instance segmentation. Specifically, we propose a Background Weaken Module (BWM) to weaken the attention of our model to the nucleus background by integrating low-level features into high-level features. To improve the robustness of the model to the hard-boundary of nuclei, we further design a Gradient-based boundary adaptive Strategy (GS) which generates boundary-weakened data for model training in an adversarial manner. We conduct extensive experiments on MoNuSeg and CPM-17 datasets, and experimental results show that our HBANet outperforms the state-of-the-art methods.",

keywords = "Adaptive strategy, Deep learning, Nuclei instance segmentation",

author = "Yalu Cheng and Pengchong Qiao and Hongliang He and Guoli Song and Jie Chen",

note = "Publisher Copyright: {\textcopyright} 2021 ACM.; 3rd ACM International Conference on Multimedia in Asia, MMAsia 2021 ; Conference date: 01-12-2021 Through 03-12-2021",

year = "2021",

month = dec,

day = "1",

doi = "10.1145/3469877.3490602",

language = "English",

series = "ACM International Conference Proceeding Series",

publisher = "Association for Computing Machinery",

booktitle = "Proceedings of the 3rd ACM International Conference on Multimedia in Asia, MMAsia 2021",

}

Cheng, Y, Qiao, P, He, H, Song, G & Chen, J 2021, Hard-Boundary Attention Network for Nuclei Instance Segmentation. in Proceedings of the 3rd ACM International Conference on Multimedia in Asia, MMAsia 2021., 22, ACM International Conference Proceeding Series, Association for Computing Machinery, 3rd ACM International Conference on Multimedia in Asia, MMAsia 2021, Virtual, Online, Australia, 1/12/21. https://doi.org/10.1145/3469877.3490602

Hard-Boundary Attention Network for Nuclei Instance Segmentation. / Cheng, Yalu; Qiao, Pengchong; He, Hongliang et al.
Proceedings of the 3rd ACM International Conference on Multimedia in Asia, MMAsia 2021. Association for Computing Machinery, 2021. 22 (ACM International Conference Proceeding Series).

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

TY - GEN

T1 - Hard-Boundary Attention Network for Nuclei Instance Segmentation

AU - Cheng, Yalu

AU - Qiao, Pengchong

AU - He, Hongliang

AU - Song, Guoli

AU - Chen, Jie

PY - 2021/12/1

Y1 - 2021/12/1

N2 - Image segmentation plays an important role in medical image analysis, and accurate segmentation of nuclei is especially crucial to clinical diagnosis. However, existing methods fail to segment dense nuclei due to the hard-boundary which has similar texture to nuclear inside. To this end, we propose a Hard-Boundary Attention Network (HBANet) for nuclei instance segmentation. Specifically, we propose a Background Weaken Module (BWM) to weaken the attention of our model to the nucleus background by integrating low-level features into high-level features. To improve the robustness of the model to the hard-boundary of nuclei, we further design a Gradient-based boundary adaptive Strategy (GS) which generates boundary-weakened data for model training in an adversarial manner. We conduct extensive experiments on MoNuSeg and CPM-17 datasets, and experimental results show that our HBANet outperforms the state-of-the-art methods.

AB - Image segmentation plays an important role in medical image analysis, and accurate segmentation of nuclei is especially crucial to clinical diagnosis. However, existing methods fail to segment dense nuclei due to the hard-boundary which has similar texture to nuclear inside. To this end, we propose a Hard-Boundary Attention Network (HBANet) for nuclei instance segmentation. Specifically, we propose a Background Weaken Module (BWM) to weaken the attention of our model to the nucleus background by integrating low-level features into high-level features. To improve the robustness of the model to the hard-boundary of nuclei, we further design a Gradient-based boundary adaptive Strategy (GS) which generates boundary-weakened data for model training in an adversarial manner. We conduct extensive experiments on MoNuSeg and CPM-17 datasets, and experimental results show that our HBANet outperforms the state-of-the-art methods.

KW - Adaptive strategy

KW - Deep learning

KW - Nuclei instance segmentation

UR - https://www.scopus.com/pages/publications/85123051772

U2 - 10.1145/3469877.3490602

DO - 10.1145/3469877.3490602

M3 - Conference contribution

AN - SCOPUS:85123051772

T3 - ACM International Conference Proceeding Series

BT - Proceedings of the 3rd ACM International Conference on Multimedia in Asia, MMAsia 2021

PB - Association for Computing Machinery

T2 - 3rd ACM International Conference on Multimedia in Asia, MMAsia 2021

Y2 - 1 December 2021 through 3 December 2021

ER -

Hard-Boundary Attention Network for Nuclei Instance Segmentation

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this