Segmenting neuronal structure in 3D optical microscope images via knowledge distillation with teacher-student network

Heng Wang; Donghao Zhang; Yang Song; Siqi Liu; Yue Wang; Dagan Feng; Hanchuan Peng; Weidong Cai

doi:10.1109/ISBI.2019.8759326

Segmenting neuronal structure in 3D optical microscope images via knowledge distillation with teacher-student network

Heng Wang, Donghao Zhang, Yang Song, Siqi Liu, Yue Wang, Dagan Feng, Hanchuan Peng, Weidong Cai

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

21 Citations (Scopus)

Abstract

Three-dimensional (3D) volumetric neural image segmentation is crucial to reconstructing accurate neuron structures. However, due to the structural complexity of neurons and the diverse imaging qualities of the microscopes, it is challenging to achieve both accuracy and efficiency. In this paper, we propose a teacher-student learning framework for fast neuron segmentation. The segmentation inference is performed using a light-weighted student network which benefits from knowledge distillation of a teacher network with a higher capacity. Evaluated on the Janelia dataset from the BigNeuron project, our proposed framework achieves competitive performance for segmentation accuracy while reducing the computational cost to facilitate large-scale processing.

Original language	English
Title of host publication	Proceedings of 2019 IEEE International Symposium on Biomedical Imaging
Place of Publication	USA
Publisher	IEEE Computer Society
Pages	228-231
Number of pages	4
ISBN (Electronic)	9781538636411
ISBN (Print)	9781538636404
DOIs	https://doi.org/10.1109/ISBI.2019.8759326
Publication status	Published - 2019
Externally published	Yes
Event	IEEE International Symposium on Biomedical Imaging (ISBI) 2019 - Hilton Molino Stucky, Venice, Italy Duration: 8 Apr 2019 → 11 Apr 2019 Conference number: 16th https://ieeexplore.ieee.org/xpl/conhome/8754684/proceeding (Proceedings)

Publication series

Name	Proceedings - International Symposium on Biomedical Imaging
Volume	2019-April
ISSN (Print)	1945-7928
ISSN (Electronic)	1945-8452

Conference

Conference	IEEE International Symposium on Biomedical Imaging (ISBI) 2019
Abbreviated title	ISBI 2019
Country/Territory	Italy
City	Venice
Period	8/04/19 → 11/04/19
Internet address	https://ieeexplore.ieee.org/xpl/conhome/8754684/proceeding (Proceedings)

Keywords

Bigneuron
Knowledge distillation
Neuronal image segmentation
Teacher-student network

Access to Document

10.1109/ISBI.2019.8759326

Cite this

Wang, H., Zhang, D., Song, Y., Liu, S., Wang, Y., Feng, D., Peng, H., & Cai, W. (2019). Segmenting neuronal structure in 3D optical microscope images via knowledge distillation with teacher-student network. In Proceedings of 2019 IEEE International Symposium on Biomedical Imaging (pp. 228-231). [8759326] (Proceedings - International Symposium on Biomedical Imaging; Vol. 2019-April). IEEE Computer Society. https://doi.org/10.1109/ISBI.2019.8759326

@inproceedings{aed245de26924614b37b78e1f6b935b1,

title = "Segmenting neuronal structure in 3D optical microscope images via knowledge distillation with teacher-student network",

abstract = "Three-dimensional (3D) volumetric neural image segmentation is crucial to reconstructing accurate neuron structures. However, due to the structural complexity of neurons and the diverse imaging qualities of the microscopes, it is challenging to achieve both accuracy and efficiency. In this paper, we propose a teacher-student learning framework for fast neuron segmentation. The segmentation inference is performed using a light-weighted student network which benefits from knowledge distillation of a teacher network with a higher capacity. Evaluated on the Janelia dataset from the BigNeuron project, our proposed framework achieves competitive performance for segmentation accuracy while reducing the computational cost to facilitate large-scale processing.",

keywords = "Bigneuron, Knowledge distillation, Neuronal image segmentation, Teacher-student network",

author = "Heng Wang and Donghao Zhang and Yang Song and Siqi Liu and Yue Wang and Dagan Feng and Hanchuan Peng and Weidong Cai",

note = "Publisher Copyright: {\textcopyright} 2019 IEEE.; IEEE International Symposium on Biomedical Imaging (ISBI) 2019, ISBI 2019 ; Conference date: 08-04-2019 Through 11-04-2019",

year = "2019",

doi = "10.1109/ISBI.2019.8759326",

language = "English",

isbn = "9781538636404",

series = "Proceedings - International Symposium on Biomedical Imaging",

publisher = "IEEE Computer Society",

pages = "228--231",

booktitle = "Proceedings of 2019 IEEE International Symposium on Biomedical Imaging",

address = "United States of America",

url = "https://ieeexplore.ieee.org/xpl/conhome/8754684/proceeding",

}

Wang, H, Zhang, D, Song, Y, Liu, S, Wang, Y, Feng, D, Peng, H & Cai, W 2019, Segmenting neuronal structure in 3D optical microscope images via knowledge distillation with teacher-student network. in Proceedings of 2019 IEEE International Symposium on Biomedical Imaging., 8759326, Proceedings - International Symposium on Biomedical Imaging, vol. 2019-April, IEEE Computer Society, USA, pp. 228-231, IEEE International Symposium on Biomedical Imaging (ISBI) 2019, Venice, Italy, 8/04/19. https://doi.org/10.1109/ISBI.2019.8759326

Segmenting neuronal structure in 3D optical microscope images via knowledge distillation with teacher-student network. / Wang, Heng; Zhang, Donghao; Song, Yang et al.

Proceedings of 2019 IEEE International Symposium on Biomedical Imaging. USA : IEEE Computer Society, 2019. p. 228-231 8759326 (Proceedings - International Symposium on Biomedical Imaging; Vol. 2019-April).

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

TY - GEN

T1 - Segmenting neuronal structure in 3D optical microscope images via knowledge distillation with teacher-student network

AU - Wang, Heng

AU - Zhang, Donghao

AU - Song, Yang

AU - Liu, Siqi

AU - Wang, Yue

AU - Feng, Dagan

AU - Peng, Hanchuan

AU - Cai, Weidong

N1 - Conference code: 16th

PY - 2019

Y1 - 2019

N2 - Three-dimensional (3D) volumetric neural image segmentation is crucial to reconstructing accurate neuron structures. However, due to the structural complexity of neurons and the diverse imaging qualities of the microscopes, it is challenging to achieve both accuracy and efficiency. In this paper, we propose a teacher-student learning framework for fast neuron segmentation. The segmentation inference is performed using a light-weighted student network which benefits from knowledge distillation of a teacher network with a higher capacity. Evaluated on the Janelia dataset from the BigNeuron project, our proposed framework achieves competitive performance for segmentation accuracy while reducing the computational cost to facilitate large-scale processing.

AB - Three-dimensional (3D) volumetric neural image segmentation is crucial to reconstructing accurate neuron structures. However, due to the structural complexity of neurons and the diverse imaging qualities of the microscopes, it is challenging to achieve both accuracy and efficiency. In this paper, we propose a teacher-student learning framework for fast neuron segmentation. The segmentation inference is performed using a light-weighted student network which benefits from knowledge distillation of a teacher network with a higher capacity. Evaluated on the Janelia dataset from the BigNeuron project, our proposed framework achieves competitive performance for segmentation accuracy while reducing the computational cost to facilitate large-scale processing.

KW - Bigneuron

KW - Knowledge distillation

KW - Neuronal image segmentation

KW - Teacher-student network

UR - http://www.scopus.com/inward/record.url?scp=85073910719&partnerID=8YFLogxK

U2 - 10.1109/ISBI.2019.8759326

DO - 10.1109/ISBI.2019.8759326

M3 - Conference Paper

AN - SCOPUS:85073910719

SN - 9781538636404

T3 - Proceedings - International Symposium on Biomedical Imaging

SP - 228

EP - 231

BT - Proceedings of 2019 IEEE International Symposium on Biomedical Imaging

PB - IEEE Computer Society

CY - USA

T2 - IEEE International Symposium on Biomedical Imaging (ISBI) 2019

Y2 - 8 April 2019 through 11 April 2019

ER -

Wang H, Zhang D, Song Y, Liu S, Wang Y, Feng D et al. Segmenting neuronal structure in 3D optical microscope images via knowledge distillation with teacher-student network. In Proceedings of 2019 IEEE International Symposium on Biomedical Imaging. USA: IEEE Computer Society. 2019. p. 228-231. 8759326. (Proceedings - International Symposium on Biomedical Imaging). https://doi.org/10.1109/ISBI.2019.8759326

Segmenting neuronal structure in 3D optical microscope images via knowledge distillation with teacher-student network

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Cite this