Table 5.

Results without CycleGAN.

Network	Accuracy (%)	Precision (%)	Recall (%)	F1-score (%)	AUC (%)	ECE
Densenet-121	88.05 ± 3.81	80.94 ± 5.55	93.18 ± 3.70	87.36 ± 3.63	96.71 ± 2.80	0.118 ± 0.130
EfficientNet-B3	94.69 ± 2.04	92.55 ± 3.40	96.77 ± 1.57	94.09 ± 2.19	99.03 ± 0.66	0.067 ± 0.012
Resnet-50	94.69 ± 1.15	90.31 ± 2.21	98.74 ± 0.48	94.25 ± 1.15	99.43 ± 0.20	0.060 ± 0.025
ResNeSt-50	96.30 ± 2.31	93.96 ± 3.00	98.02 ± 2.43	95.97 ± 2.53	99.60 ± 1.18	0.020 ± 0.007
ViT	99.60 ± 0.79	99.46 ± 1.39	99.64 ± 0.38	99.55 ± 0.88	99.99 ± 0.10	0.014 ± 0.007