Auto-classification of Retinal Diseases in the Limit of Sparse Data Using a Two-Streams Machine Learning Model

Abstract
Automatic clinical diagnosis of retinal diseases has emerged as a promising approach to facilitate discovery in areas with limited access to specialists. Based on the fact that fundus structure and vascular disorders are the main characteristics of retinal diseases, we propose a novel visual-assisted diagnosis hybrid model mixing the support vector machine (SVM) and deep neural networks (DNNs). Furthermore, we present a new clinical retina labels collection sorted by the professional ophthalmologist from the educational project Retina Image Bank, called EyeNet, for ophthalmology incorporating 52 retina diseases classes. Using EyeNet, our model achieves 90.40% diagnosis accuracy, and the model performance is comparable to the professional ophthalmologists (https://github.com/huckiyang/EyeNet2).

Citation
Huck Yang, C.-H., Liu, F., Huang, J.-H., Tian, M., I-Hung Lin, M. D., Liu, Y. C., … Tegnèr, J. (2019). Auto-classification of Retinal Diseases in the Limit of Sparse Data Using a Two-Streams Machine Learning Model. Lecture Notes in Computer Science, 323–338. doi:10.1007/978-3-030-21074-8_28

Publisher
Springer Nature

Conference/Event Name
14th Asian Conference on Computer Vision, ACCV 2018

DOI
10.1007/978-3-030-21074-8_28

arXiv
1808.05754

Additional Links
http://link.springer.com/10.1007/978-3-030-21074-8_28

Relations
Is Supplemented By:
  • [Software]
    Title: huckiyang/EyeNet2: ACCV 18 - Auto-Classification of Retinal Diseases in the Limit of Sparse Data Using a Two-Streams Machine Learning Model. Publication Date: 2018-07-02. github: huckiyang/EyeNet2 Handle: 10754/668078

Permanent link to this record