Classification of Diabetic Retinopathy Severity in Fundus Images Using the Vision Transformer and Residual Attention

Table 7

The results of different ablation experiments in the DDR dataset. Sensitivity, specificity, accuracy, and AUC are average values, respectively.

Method	Sensitivity	Specificity	Accuracy	AUC
ResNet50 + CSRA	0.7941	0.7846	0.7940	0.8405
ResNet101+CSRA	0.8120	0.8242	0.8130	0.8903
ViT + MLP	0.7929	0.7815	0.7980	0.8647
ViT + CSRA	0.8140	0.8245	0.8235	0.9018

The bold font represents the best performance in each column. The italic font represents the second-best performance in each column.