Open Access

ARTICLE

Scheme Based on Multi-Level Patch Attention and Lesion Localization for Diabetic Retinopathy Grading

by Zhuoqun Xia¹, Hangyu Hu¹, Wenjing Li^2,3, Qisheng Jiang¹, Lan Pu¹, Yicong Shu¹, Arun Kumar Sangaiah^4,5,*

1 School of Computer and Communication Engineering, Changsha University of Science and Technology, Changsha, China
2 Hunan Province People’s Hospital (The First Affiliated Hospital of Hunan Normal University), Changsha, China
3 The First Affiliated Hospital of Guangxi Medical University, Nanning, China
4 International Graduate Institute of AI, National Yunlin University of Science and Technology, Yunlin, China
5 Department of Electrical and Computer Engineering, Lebanese American University, Bayrut, Lebanon

* Corresponding Author: Arun Kumar Sangaiah. Email:

(This article belongs to the Special Issue: Smart and Secure Solutions for Medical Industry)

Computer Modeling in Engineering & Sciences 2024, 140(1), 409-430. https://doi.org/10.32604/cmes.2024.030052

Received 20 March 2023; Accepted 15 September 2023; Issue published 16 April 2024

Abstract

Early screening of diabetes retinopathy (DR) plays an important role in preventing irreversible blindness. Existing research has failed to fully explore effective DR lesion information in fundus maps. Besides, traditional attention schemes have not considered the impact of lesion type differences on grading, resulting in unreasonable extraction of important lesion features. Therefore, this paper proposes a DR diagnosis scheme that integrates a multi-level patch attention generator (MPAG) and a lesion localization module (LLM). Firstly, MPAG is used to predict patches of different sizes and generate a weighted attention map based on the prediction score and the types of lesions contained in the patches, fully considering the impact of lesion type differences on grading, solving the problem that the attention maps of lesions cannot be further refined and then adapted to the final DR diagnosis task. Secondly, the LLM generates a global attention map based on localization. Finally, the weighted attention map and global attention map are weighted with the fundus map to fully explore effective DR lesion information and increase the attention of the classification network to lesion details. This paper demonstrates the effectiveness of the proposed method through extensive experiments on the public DDR dataset, obtaining an accuracy of 0.8064.

---

Keywords

---