Source-free active domain adaptation for diabetic retinopathy grading based on ultra-wide-field fundus images

被引：1
作者：
Ran J. [1 ]
Zhang G. [2 ,4 ]
Xia F. [3 ]
Zhang X. [4 ]
Xie J. [5 ]
Zhang H. [6 ]
机构：
[1] College of Computer and Information Science, Southwest University, Chongqing
[2] School of Big Data Intelligent Diagnosis and Treatment Industry, Taiyuan University, Taiyuan
[3] Reading Academy, Nanjing University of Information Science and Technology, Nanjing
[4] College of Biomedical Engineering, Taiyuan University of Technology, Taiyuan
[5] Shanxi Eye hospital, Taiyuan
[6] College of Chemistry and Chemical Engineering, Southwest University, Chongqing
来源：
Computers in Biology and Medicine | 2024年 / 174卷
基金：
中国国家自然科学基金;
关键词：
Active learning; Diabetic retinopathy grading; Source-free domain adaptation; Ultra-wide-field fundus images;
D O I：
10.1016/j.compbiomed.2024.108418
中图分类号：
学科分类号：
摘要：
Domain adaptation (DA) is commonly employed in diabetic retinopathy (DR) grading using unannotated fundus images, allowing knowledge transfer from labeled color fundus images. Existing DAs often struggle with domain disparities, hindering DR grading performance compared to clinical diagnosis. A source-free active domain adaptation method (SFADA), which generates features of color fundus images by noise, selects valuable ultra-wide-field (UWF) fundus images through local representation matching, and adapts models using DR lesion prototypes, is proposed to upgrade DR diagnostic accuracy. Importantly, SFADA enhances data security and patient privacy by excluding source domain data. It reduces image resolution and boosts model training speed by modeling DR grade relationships directly. Experiments show SFADA significantly improves DR grading performance, increasing accuracy by 20.90% and quadratic weighted kappa by 18.63% over baseline, reaching 85.36% and 92.38%, respectively. This suggests SFADA's promise for real clinical applications. © 2024 Elsevier Ltd
引用
收藏
相关论文
共 51 条
  • [1] Wang W., Lo A.C., Diabetic retinopathy: pathophysiology and treatments, Int. J. Mol. Sci., 19, 6, (2018)
  • [2] Himasa F.I., Singhal M., Ojha A., Kumar B., Prospective for diagnosis and treatment of diabetic retinopathy, Curr. Pharm. Des., 28, 7, pp. 560-569, (2022)
  • [3] Leasher J.L., Bourne R.R., Flaxman S.R., Jonas J.B., Keeffe J., Naidoo K., Pesudovs K., Price H., White R.A., Wong T.Y., Et al., Global estimates on the number of people blind or visually impaired by diabetic retinopathy: a meta-analysis from 1990 to 2010, Diabetes Care, 39, 9, pp. 1643-1649, (2016)
  • [4] Cho N.H., Shaw J., Karuranga S., Huang Y., da Rocha Fernandes J., Ohlrogge A., Malanda B., IDF Diabetes Atlas: Global estimates of diabetes prevalence for 2017 and projections for 2045, Diabetes Res. Clin. Pract., 138, pp. 271-281, (2018)
  • [5] Wang L., Gao P., Zhang M., Huang Z., Zhang D., Deng Q., Li Y., Zhao Z., Qin X., Jin D., Et al., Prevalence and ethnic pattern of diabetes and prediabetes in China in 2013, JAMA, 317, 24, pp. 2515-2523, (2017)
  • [6] Liu Y.P., Li Z., Xu C., Li J., Liang R., Referable diabetic retinopathy identification from eye fundus images with weighted path for convolutional neural network, Artif. Intell. Med., 99, (2019)
  • [7] Echouffo-Tcheugui J., Ali M., Roglic G., Hayward R., Narayan K., Screening intervals for diabetic retinopathy and incidence of visual loss: a systematic review, Diabetic Med., 30, 11, pp. 1272-1292, (2013)
  • [8] Tung T.H., Chen S.J., Shih H.C., Chou P., Li A.F., Shyong M.P., Lee F.L., Liu J.H., Assessing the natural course of diabetic retinopathy: a population-based study in Kinmen, Taiwan, Ophthalmic Epidemiol., 13, 5, pp. 327-333, (2006)
  • [9] Stolte S., Fang R., A survey on medical image analysis in diabetic retinopathy, Med. Image Anal., 64, (2020)
  • [10] Yang Z., Tan T.E., Shao Y., Wong T.Y., Li X., Classification of diabetic retinopathy: Past, present and future, Front. Endocrinol., 13, (2022)
123456