Network Pruning for OCT Image Classification

Bing Yang; Yi Zhang; Jun Cheng; Jin Zhang; Lei Mou; Huaying Hao; Yitian Zhao; Jiang Liu

doi:10.1007/978-3-030-32956-3_15

Network Pruning for OCT Image Classification

Bing Yang, Yi Zhang, Jun Cheng, Jin Zhang, Lei Mou, Huaying Hao, Yitian Zhao, Jiang Liu

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

1 Citation (Scopus)

Abstract

Convolutional neural network (CNN) has expanded rapidly, and has been widely used in medical image classification. The large number of parameters in a neural network makes CNN models computationally expensive. This leads to slow inference speed, especially for 3D data such as optical coherence tomography (OCT) for retinal images. A volume OCT scan of retina often contains hundreds of 2D images which needs to be analyzed sequentially in a local computer with limited computational resources. We introduce network pruning to OCT images classification and propose an algorithm to prune networks. We compress the popular classification models, such as ResNet and VGG. For example, within 1% accuracy loss, we compress ResNet-18 from 44.8 MB to 69 KB and VGG-16 from 537.1 MB to 194 KB. These pruned models are much smaller and easier to deploy on the OCT devices. As for the inference speed, the pruned models are 10 to 20 times faster than original models for ResNet and VGG in CPU.

Original language	English
Title of host publication	Ophthalmic Medical Image Analysis - 6th International Workshop, OMIA 2019, Held in Conjunction with MICCAI 2019, Proceedings
Editors	Huazhu Fu, Mona K. Garvin, Tom MacGillivray, Yanwu Xu, Yalin Zheng
Publisher	Springer
Pages	121-129
Number of pages	9
ISBN (Print)	9783030329556
DOIs	https://doi.org/10.1007/978-3-030-32956-3_15
Publication status	Published - 2019
Externally published	Yes
Event	6th International Workshop on Ophthalmic Medical Image Analysis, OMIA 2019, held in conjunction with the 22nd International Conference on Medical Imaging Computing and Computer-Assisted Intervention, MICCAI 2019 - Shenzhen, China Duration: 17 Oct 2019 → 17 Oct 2019

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11855 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	6th International Workshop on Ophthalmic Medical Image Analysis, OMIA 2019, held in conjunction with the 22nd International Conference on Medical Imaging Computing and Computer-Assisted Intervention, MICCAI 2019
Country/Territory	China
City	Shenzhen
Period	17/10/19 → 17/10/19

Keywords

Classification
Network pruning
OCT

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-030-32956-3_15

Cite this

Yang, B., Zhang, Y., Cheng, J., Zhang, J., Mou, L., Hao, H., Zhao, Y., & Liu, J. (2019). Network Pruning for OCT Image Classification. In H. Fu, M. K. Garvin, T. MacGillivray, Y. Xu, & Y. Zheng (Eds.), Ophthalmic Medical Image Analysis - 6th International Workshop, OMIA 2019, Held in Conjunction with MICCAI 2019, Proceedings (pp. 121-129). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11855 LNCS). Springer. https://doi.org/10.1007/978-3-030-32956-3_15

Yang, Bing ; Zhang, Yi ; Cheng, Jun et al. / Network Pruning for OCT Image Classification. Ophthalmic Medical Image Analysis - 6th International Workshop, OMIA 2019, Held in Conjunction with MICCAI 2019, Proceedings. editor / Huazhu Fu ; Mona K. Garvin ; Tom MacGillivray ; Yanwu Xu ; Yalin Zheng. Springer, 2019. pp. 121-129 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{678c2042984a4704a769fbc578c73108,

title = "Network Pruning for OCT Image Classification",

abstract = "Convolutional neural network (CNN) has expanded rapidly, and has been widely used in medical image classification. The large number of parameters in a neural network makes CNN models computationally expensive. This leads to slow inference speed, especially for 3D data such as optical coherence tomography (OCT) for retinal images. A volume OCT scan of retina often contains hundreds of 2D images which needs to be analyzed sequentially in a local computer with limited computational resources. We introduce network pruning to OCT images classification and propose an algorithm to prune networks. We compress the popular classification models, such as ResNet and VGG. For example, within 1% accuracy loss, we compress ResNet-18 from 44.8 MB to 69 KB and VGG-16 from 537.1 MB to 194 KB. These pruned models are much smaller and easier to deploy on the OCT devices. As for the inference speed, the pruned models are 10 to 20 times faster than original models for ResNet and VGG in CPU.",

keywords = "Classification, Network pruning, OCT",

author = "Bing Yang and Yi Zhang and Jun Cheng and Jin Zhang and Lei Mou and Huaying Hao and Yitian Zhao and Jiang Liu",

note = "Publisher Copyright: {\textcopyright} Springer Nature Switzerland AG 2019.; 6th International Workshop on Ophthalmic Medical Image Analysis, OMIA 2019, held in conjunction with the 22nd International Conference on Medical Imaging Computing and Computer-Assisted Intervention, MICCAI 2019 ; Conference date: 17-10-2019 Through 17-10-2019",

year = "2019",

doi = "10.1007/978-3-030-32956-3_15",

language = "English",

isbn = "9783030329556",

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

publisher = "Springer",

pages = "121--129",

editor = "Huazhu Fu and Garvin, {Mona K.} and Tom MacGillivray and Yanwu Xu and Yalin Zheng",

booktitle = "Ophthalmic Medical Image Analysis - 6th International Workshop, OMIA 2019, Held in Conjunction with MICCAI 2019, Proceedings",

}

Yang, B, Zhang, Y, Cheng, J, Zhang, J, Mou, L, Hao, H, Zhao, Y & Liu, J 2019, Network Pruning for OCT Image Classification. in H Fu, MK Garvin, T MacGillivray, Y Xu & Y Zheng (eds), Ophthalmic Medical Image Analysis - 6th International Workshop, OMIA 2019, Held in Conjunction with MICCAI 2019, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11855 LNCS, Springer, pp. 121-129, 6th International Workshop on Ophthalmic Medical Image Analysis, OMIA 2019, held in conjunction with the 22nd International Conference on Medical Imaging Computing and Computer-Assisted Intervention, MICCAI 2019, Shenzhen, China, 17/10/19. https://doi.org/10.1007/978-3-030-32956-3_15

Network Pruning for OCT Image Classification. / Yang, Bing; Zhang, Yi; Cheng, Jun et al.
Ophthalmic Medical Image Analysis - 6th International Workshop, OMIA 2019, Held in Conjunction with MICCAI 2019, Proceedings. ed. / Huazhu Fu; Mona K. Garvin; Tom MacGillivray; Yanwu Xu; Yalin Zheng. Springer, 2019. p. 121-129 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11855 LNCS).

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

TY - GEN

T1 - Network Pruning for OCT Image Classification

AU - Yang, Bing

AU - Zhang, Yi

AU - Cheng, Jun

AU - Zhang, Jin

AU - Mou, Lei

AU - Hao, Huaying

AU - Zhao, Yitian

AU - Liu, Jiang

N1 - Publisher Copyright: © Springer Nature Switzerland AG 2019.

PY - 2019

Y1 - 2019

N2 - Convolutional neural network (CNN) has expanded rapidly, and has been widely used in medical image classification. The large number of parameters in a neural network makes CNN models computationally expensive. This leads to slow inference speed, especially for 3D data such as optical coherence tomography (OCT) for retinal images. A volume OCT scan of retina often contains hundreds of 2D images which needs to be analyzed sequentially in a local computer with limited computational resources. We introduce network pruning to OCT images classification and propose an algorithm to prune networks. We compress the popular classification models, such as ResNet and VGG. For example, within 1% accuracy loss, we compress ResNet-18 from 44.8 MB to 69 KB and VGG-16 from 537.1 MB to 194 KB. These pruned models are much smaller and easier to deploy on the OCT devices. As for the inference speed, the pruned models are 10 to 20 times faster than original models for ResNet and VGG in CPU.

AB - Convolutional neural network (CNN) has expanded rapidly, and has been widely used in medical image classification. The large number of parameters in a neural network makes CNN models computationally expensive. This leads to slow inference speed, especially for 3D data such as optical coherence tomography (OCT) for retinal images. A volume OCT scan of retina often contains hundreds of 2D images which needs to be analyzed sequentially in a local computer with limited computational resources. We introduce network pruning to OCT images classification and propose an algorithm to prune networks. We compress the popular classification models, such as ResNet and VGG. For example, within 1% accuracy loss, we compress ResNet-18 from 44.8 MB to 69 KB and VGG-16 from 537.1 MB to 194 KB. These pruned models are much smaller and easier to deploy on the OCT devices. As for the inference speed, the pruned models are 10 to 20 times faster than original models for ResNet and VGG in CPU.

KW - Classification

KW - Network pruning

KW - OCT

UR - http://www.scopus.com/inward/record.url?scp=85075674635&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-32956-3_15

DO - 10.1007/978-3-030-32956-3_15

M3 - Conference contribution

AN - SCOPUS:85075674635

SN - 9783030329556

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

SP - 121

EP - 129

BT - Ophthalmic Medical Image Analysis - 6th International Workshop, OMIA 2019, Held in Conjunction with MICCAI 2019, Proceedings

A2 - Fu, Huazhu

A2 - Garvin, Mona K.

A2 - MacGillivray, Tom

A2 - Xu, Yanwu

A2 - Zheng, Yalin

PB - Springer

T2 - 6th International Workshop on Ophthalmic Medical Image Analysis, OMIA 2019, held in conjunction with the 22nd International Conference on Medical Imaging Computing and Computer-Assisted Intervention, MICCAI 2019

Y2 - 17 October 2019 through 17 October 2019

ER -

Yang B, Zhang Y, Cheng J, Zhang J, Mou L, Hao H et al. Network Pruning for OCT Image Classification. In Fu H, Garvin MK, MacGillivray T, Xu Y, Zheng Y, editors, Ophthalmic Medical Image Analysis - 6th International Workshop, OMIA 2019, Held in Conjunction with MICCAI 2019, Proceedings. Springer. 2019. p. 121-129. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-32956-3_15

Network Pruning for OCT Image Classification

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this