[Retracted] A Deep Learning Framework for Earlier Prediction of Diabetic Retinopathy from Fundus Photographs

Table 6

Effectiveness on replica testing sets, comparable to the actual research’s findings.

Testing datasets	High specificity	High sensitivity	Area under the ROC curve scores
KaggleEyePACS testing (actual EyePACS)	94.7 (92.4%)	95.7 (99.6)%	0.972 (1.095)
	95.1 (99.2)%	95.8 (95.5)%