GasHis-Transformer: A multi-scale visual transformer approach for gastric histopathological image detection

被引：143

作者：

Chen, Haoyuan ^{[1
]}

Li, Chen ^{[1
]}

Wang, Ge ^{[2
]}

Li, Xiaoyan ^{[4
]}

Rahaman, Md Mamunur ^{[1
,3
]}

Sun, Hongzan ^{[4
]}

Hu, Weiming ^{[1
]}

Li, Yixin ^{[1
]}

Liu, Wanli ^{[1
]}

Sun, Changhao ^{[1
,5
]}

Ai, Shiliang ^{[1
]}

Grzegorzek, Marcin ^{[6
]}

机构：

[1] Northeastern Univ, Coll Med & Biol Informat Engn, Microscop Image & Med Image Anal Grp, Shenyang, Peoples R China

[2] Rensselaer Polytech Inst, Biomed Imaging Ctr, Dept Biomed Engn, Troy, NY 12180 USA

[3] Univ New South Wales, Sch Comp Sci & Engn, Sydney, NSW, Australia

[4] China Med Univ, Shenyang, Peoples R China

[5] Chinese Acad Sci, Shenyang Inst Automat, Shenyang, Peoples R China

[6] Univ Lubeck, Inst Med Informat, Lubeck, Germany

来源：

PATTERN RECOGNITION | 2022年 / 130卷

基金：

中国国家自然科学基金;

关键词：

Gastric histropathological image; Multi-scale visual transformer; Image detection;

D O I：

10.1016/j.patcog.2022.108827

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

In this paper, a multi-scale visual transformer model, referred as GasHis-Transformer, is proposed for Gastric Histopathological Image Detection (GHID), which enables the automatic global detection of gastric cancer images. GasHis-Transformer model consists of two key modules designed to extract global and local information using a position-encoded transformer model and a convolutional neural network with local convolution, respectively. A publicly available hematoxylin and eosin (H&E) stained gastric histopathological image dataset is used in the experiment. Furthermore, a Dropconnect based lightweight network is proposed to reduce the model size and training time of GasHis-Transformer for clinical applications with improved confidence. Moreover, a series of contrast and extended experiments verify the robustness, extensibility and stability of GasHis-Transformer. In conclusion, GasHis-Transformer demonstrates high global detection performance and shows its significant potential in GHID task. (C) 2022 Elsevier Ltd. All rights reserved.

引用

页数：13

共 40 条

[1] A State-of-the-Art Review for Gastric Histopathology Image Analysis Approaches and Future Development [J].

Ai, Shiliang ;

Li, Chen ;

Li, Xiaoyan ;

Jiang, Tao ;

Grzegorzek, Marcin ;

Sun, Changhao ;

Rahaman, Md Mamunur ;

Zhang, Jinghua ;

Yao, Yudong ;

Li, Hong .

BIOMED RESEARCH INTERNATIONAL, 2021, 2021

[2] Attention Augmented Convolutional Networks [J].

Bello, Irwan ;

Zoph, Barret ;

Vaswani, Ashish ;

Shlens, Jonathon ;

Le, Quoc V. .

2019 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2019), 2019, :3285-3294

[3] GCNet: Non-local Networks Meet Squeeze-Excitation Networks and Beyond [J].

Cao, Yue ;

Xu, Jiarui ;

Lin, Stephen ;

Wei, Fangyun ;

Hu, Han .

2019 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION WORKSHOPS (ICCVW), 2019, :1971-1980

[4] Xception: Deep Learning with Depthwise Separable Convolutions [J].

Chollet, Francois .

30TH IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2017), 2017, :1800-1807

[5] A comprehensive survey on model compression and acceleration [J].

Choudhary, Tejalal ;

Mishra, Vipul ;

Goswami, Anurag ;

Sarangapani, Jagannathan .

ARTIFICIAL INTELLIGENCE REVIEW, 2020, 53 (07) :5113-5155

[6] TransMed: Transformers Advance Multi-Modal Medical Image Classification [J].

Dai, Yin ;

Gao, Yifan ;

Liu, Fayu .

DIAGNOSTICS, 2021, 11 (08)

[7]

Dosovitskiy A., 2021, P 9 INT C LEARN REPR

[8] A black-box adversarial attack strategy with adjustable sparsity and generalizability for deep image classifiers [J].

Ghosh, Arka ;

Mullick, Sankha Subhra ;

Datta, Shounak ;

Das, Swagatam ;

Das, Asit Kr ;

Mallipeddi, Rammohan .

PATTERN RECOGNITION, 2022, 122

[9]

Goodfellow I. J., 2015, ICLR

[10]

Graham B, 2021, Arxiv, DOI [arXiv:2104.01136, 10.48550/arXiv.2104.01136]

← 1 2 3 4 →