A study on construction of an eye disease classification and diagnosis model by utilizing resnet deep neural network
Objective The purpose of this study was to develop a diagnostic model based on the ResNet50 deep neural network for the classification and diagnosis of glaucoma,cataracts,and diabetic retinopathy.Methods The ResNet50 model was trained on 4 217 fundus images from the eye disease classification data and was evaluated using 3 000 epochs.The model parameters with the highest accuracy on the test set were selected for the diagnostic model.Performance was evaluated using metrics such as confusion matrix,accuracy,recall,F1 score,and the area under the receiver operating characteristic curve(AUC).Results The diagnostic model achieved an accuracy of 94.25％on the test set,with a precision of 94.42％,a recall of 94.25％,an F1 score of 94.23％,and a mean AUC of 0.9856.These results indicate that the diagnostic model has high accuracy for the classification and diagnosis of glaucoma,cataracts,and diabetic reti-nopathy.Conclusion The diagnostic model based on the ResNet50 deep neural network is considered to have good per-formance and can be used for the prediction of glaucoma,cataracts,and diabetic retinopathy,providing valuable refer-ence for clinical classification and diagnosis.
Deep learning networkResNet50Classification diagnosis modelGlaucomacataractDiabetic retinopathy
NETL
NSTL
万方数据
