Skip to main content

Table 2 The performance of the two-stage method for no-DR and DR categories

From: General deep learning model for detecting diabetic retinopathy

DataBase

Accuracy (%)

Sensitivity (%)

Specificity (%)

Precision (%)

Train & validation set

 EyePACs

85.00

84.89

85.29

94.17

Test set

 DIARETDB0

85.38

100

82.73

51.28

 DIARETDB1

84.27

100

83.53

22.22

 e-ophtha

85.75

89.18

81.03

86.59

 MESSIDOR

86.73

91.39

83.12

80.74

 DRIVER

92.50

93.94

85.71

96.88