Efficient pyramid channel attention network for pathological myopia recognition with pretraining-and-finetuning

doi:10.1016/j.artmed.2024.102926

	Efficient pyramid channel attention network for pathological myopia recognition with pretraining-and-finetuning
	Zhang, Xiaoqing 1,2,3,4; Zhao, Jilu 1,2; Li, Yan 5,6; Wu, Hao 5,6; Zhou, Xiangtian 5,6,7; Liu, Jiang 1,2,5,8
	2024-08-01
发表期刊	ARTIFICIAL INTELLIGENCE IN MEDICINE
ISSN	0933-3657
卷号	154 页码:12
摘要	Pathological myopia (PM) is the leading ocular disease for impaired vision worldwide. Clinically, the characteristics of pathology distribution in PM are global-local on the fundus image, which plays a significant role in assisting clinicians in diagnosing PM. However, most existing deep neural networks focused on designing complex architectures but rarely explored the pathology distribution prior of PM. To tackle this issue, we propose an efficient pyramid channel attention (EPCA) module, which fully leverages the potential of the clinical pathology prior of PM with pyramid pooling and multi-scale context fusion. Then, we construct EPCANet for automatic PM recognition based on fundus images by stacking a sequence of EPCA modules. Moreover, motivated by the recent pretraining-and-finetuning paradigm, we attempt to adapt pre-trained natural image models for PM recognition by freezing them and treating the EPCA and other attention modules as adapters. In addition, we construct a PM recognition benchmark termed PM-fundus by collecting fundus images of PM from publicly available datasets. The comprehensive experiments demonstrate the superiority of EPCA-Net over state-of-the-art methods in the PM recognition task. For example, EPCA-Net achieves 97.56% accuracy and outperforms ViT by 2.85% accuracy on the PM-fundus dataset. The results also show that our method based on the pretraining-and-finetuning paradigm achieves competitive performance through comparisons to part of previous methods based on traditional fine-tuning paradigm with fewer tunable parameters, which has the potential to leverage more natural image foundation models to address the PM recognition task in limited medical data regime.
关键词	Pathological myopia recognition Efficient pyramid channel attention Adapter Pretraining-and-finetuning
DOI	10.1016/j.artmed.2024.102926
收录类别	SCI
语种	英语
资助项目	National Natural Science Foun-dation of China[82272086] ; Leading Goose Program of Zhejiang, China[2023C03079] ; Shenzhen Natural Science Fund, China[JCYJ20200109140820699]
WOS研究方向	Computer Science ; Engineering ; Medical Informatics
WOS类目	Computer Science, Artificial Intelligence ; Engineering, Biomedical ; Medical Informatics
WOS记录号	WOS:001265152300001
出版者	ELSEVIER
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/39851
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Liu, Jiang
作者单位	1.Southern Univ Sci & Technol, Res Inst Trustworthy Autonomous Syst, Shenzhen 518055, Peoples R China 2.Southern Univ Sci & Technol, Dept Comp Sci & Engn, Shenzhen 518055, Peoples R China 3.Chinese Acad Sci, Ctr High Performance Comp, Shenzhen 518055, Peoples R China 4.Chinese Acad Sci, Shenzhen Inst Adv Technol, Shenzhen Key Lab Intelligent Bioinformat, Shenzhen 518055, Peoples R China 5.Wenzhou Med Univ, Eye Hosp, Natl Clin Res Ctr Ocular Dis, Wenzhou 325027, Peoples R China 6.Wenzhou Med Univ, Eye Hosp, State Key Lab Ophthalmol Optometry & Visual Sci, Wenzhou 325027, Peoples R China 7.Chinese Acad Med Sci, Res Unit Myopia Basic Res & Clin Prevent & Contro, Wenzhou 325027, Peoples R China 8.Singapore Eye Res Inst, Singapore 169856, Singapore
推荐引用方式 GB/T 7714	Zhang, Xiaoqing,Zhao, Jilu,Li, Yan,et al. Efficient pyramid channel attention network for pathological myopia recognition with pretraining-and-finetuning[J]. ARTIFICIAL INTELLIGENCE IN MEDICINE,2024,154:12.
APA	Zhang, Xiaoqing,Zhao, Jilu,Li, Yan,Wu, Hao,Zhou, Xiangtian,&Liu, Jiang.(2024).Efficient pyramid channel attention network for pathological myopia recognition with pretraining-and-finetuning.ARTIFICIAL INTELLIGENCE IN MEDICINE,154,12.
MLA	Zhang, Xiaoqing,et al."Efficient pyramid channel attention network for pathological myopia recognition with pretraining-and-finetuning".ARTIFICIAL INTELLIGENCE IN MEDICINE 154(2024):12.