Table 6.

Performance comparison with state-of-the-art methods using MESSIDOR-2.

Model	ACC	AUC	SE	SP
Gulshan et al.17	—	0.961	0.939	0.961
Voets et al.20	—	0.800	0.900	0.760
Rakhlin et al.54	—	0.970	0.990	0.710
Pratt et al.55	0.750	—	0.300	0.950
Gargeya et al.56	—	0.940	0.930	0.870
Ours	0.962	0.979	0.967	0.891

Note: The values in bold show the best performing model results.