Study on detection method of diabetes retinopathy based on convolution neural network
Conference: EEI 2022 - 4th International Conference on Electronic Engineering and Informatics
06/24/2022 - 06/26/2022 at Guiyang, China
Proceedings: EEI 2022
Pages: 5Language: englishTyp: PDF
Authors:
Pan, Sen; Liu, Fenglian (School of Computer Science and Engineering, Tianjin University of Technology, Tianjing, China)
Wang, Riwei (Wenzhou University of Technology, Wenzhou, Zhejiang, China)
Abstract:
Diabetic retinopathy (DR) is a common eye disease caused by diabetes and the leading cause of blindness in diabetic patients. In order to conveniently judge whether a large number of diabetic patients suffer from DR, a computer-aided method for automatic diagnosis of DR is proposed. Different from the previous use of fundus photography to diagnose, innovatively uses the conjunctival image of the eye as the data source, and uses a new convolutional neural network ConvNeXt to train the data set. The accuracy of binary classification task reached 94.8%. In addition, the sensitivity and specificity were 93.3% and 96.4% respectively. The model can be used for DR screening.