共 68 条
Cross-Scale Fuzzy Holistic Attention Network for Diabetic Retinopathy Grading From Fundus Images
被引:2
作者:
Lin, Zhijie
[1
,2
]
He, Zhaoshui
[1
,2
]
Wang, Xu
[3
]
Su, Wenqing
[1
,4
]
Tan, Ji
[1
,4
]
Deng, Yamei
[5
]
Xie, Shengli
[1
,4
]
机构:
[1] Guangdong Univ Technol, Sch Automat, Guangzhou 510006, Peoples R China
[2] Guangdong Prov Key Lab Intelligent Syst & Optimiza, Guangzhou 510006, Peoples R China
[3] Guangdong Mech & Elect Polytech, Sch Elect & Commun, Guangzhou 510550, Peoples R China
[4] Minist Educ, Key Lab IoT Intelligent Informat Proc & Syst Integ, Guangzhou 510006, Peoples R China
[5] Guangzhou Med Univ, Affiliated Hosp 3, Dept Radiol, Guangzhou 510150, Peoples R China
来源:
IEEE TRANSACTIONS ON EMERGING TOPICS IN COMPUTATIONAL INTELLIGENCE
|
2025年
/
9卷
/
03期
基金:
中国国家自然科学基金;
关键词:
Lesions;
Feature extraction;
Retina;
Deep learning;
Uncertainty;
Solid modeling;
Medical diagnostic imaging;
Support vector machines;
Visual impairment;
Interference;
Fuzzy deep learning;
diabetic retinopathy grading;
attention network;
fundus image;
computer-aided diagnosis (CAD);
D O I:
10.1109/TETCI.2025.3543361
中图分类号:
TP18 [人工智能理论];
学科分类号:
081104 ;
0812 ;
0835 ;
1405 ;
摘要:
Diabetic Retinopathy (DR) is one of the leading causes of visual impairment and blindness in diabetic patients worldwide. Accurate Computer-Aided Diagnosis (CAD) systems can aid in the early diagnosis and treatment of DR patients to reduce the risk of vision loss, but it remains challenging due to the following reasons: 1) the relatively low contrast and ambiguous boundaries between pathological lesions and normal retinal regions, and 2) the considerable diversity in lesion size and appearance. In this paper, a Cross-Scale Fuzzy Holistic Attention Network (CSFHANet) is proposed for DR grading using fundus images, and it consists of two main components: Fuzzy-Enhanced Holistic Attention (FEHA) and Fuzzy Learning-based Cross-Scale Fusion (FLCSF). FEHA is developed to adaptively recalibrate the importance of feature elements by assigning fuzzy weights across both channel and spatial domains, which can enhance the model's ability to learn the features of lesion regions while reducing the interference from irrelevant information in normal retinal regions. Then, the FLCSF module is designed to eliminate the uncertainty in fused multi-scale features derived from different branches by utilizing fuzzy membership functions, producing a more comprehensive and refined feature representation from complex DR lesions. Extensive experiments on the Messidor-2 and DDR datasets demonstrate that the proposed CSFHANet exhibits superior performance compared to state-of-the-art methods.
引用
收藏
页码:2164 / 2178
页数:15
相关论文