Table 9.

Model’s performance using Rmsprop optimizer for Study One, along with confidence interval ($\alpha =0.05$). $A_c$–accuracy, $P_r$–precision, $R_e$–recall, $F_s$–F1-score, $S_n$–sensitivity, $S_p$–specificity.

Algorithm	Training set						Testing set
	$A_c$	$P_r$	$R_e$	$F_s$	$S_n$	$S_p$	$A_c$	$P_r$	$R_e$	$F_s$	$S_n$	$S_p$
VGG16	100%	1	1	1	1	1	88% $\pm$ 7.591	0.90 ± 0.069	0.88 ± 0.076	0.87 ± 0.079	1	0.7143 ± 0.117
ResNet50	57% $\pm$ 7.420	0.32 ± 0.093	0.57 ± 0.074	0.41 ± 0.087	1	0	56% $\pm$ 14.53	0.32 ± 0.181	0.56 ± 0.145	0.4 ± 0.170	1	0
ResNet101	62% $\pm$ 6.976	0.80 ± 0.051	0.56 ± 0.075	0.48 ± 0.082	1	0.12 ± 0.106	56% $\pm$ 14.536	0.28 ± 0.186	0.5 ± 0.155	0.36 ± 0.175	1	0
Xception	100%	1	1	1	1	1	94%$\pm$5.36	0.94 ± 0.054	0.94 ± 0.054	0.94 ± 0.054	1	0.8571 ± 0.083
EfficientNetB0	57% $\pm$ 7.420	0.32 ± 0.093	0.57 ± 0.074	0.41 ± 0.087	1	0	56% $\pm$ 14.53	0.32 ± 0.181	0.56 ± 0.145	0.4 ± 0.170	1	0
EfficientNetB7	57% $\pm$ 7.420	0.32 ± 0.093	0.57 ± 0.074	0.41 ± 0.087	1	0	56% $\pm$ 14.53	0.32 ± 0.181	0.56 ± 0.145	0.4 ± 0.170	1	0
NasNetLarge	100%	1	1	1	1	1	81% $\pm$ 9.56	0.81 ± 0.096	0.81 ± 0.096	0.81 ± 0.096	0.88 ± 0.076	0.7143 ± 0.117
EfficientNetV2M	57% $\pm$ 7.420	0.32 ± 0.093	0.57 ± 0.074	0.41 ± 0.087	1	0	56% $\pm$ 14.53	0.32 ± 0.181	0.56 ± 0.145	0.4 ± 0.170	1	0
ResNet152V2	100%	1	1	1	1	1	81%$\pm$ 9.55	0.86 ± 0.082	0.81 ± 0.096	0.8 ± 0.098	1	0.5714 ± 0.143
EfficientNetV2L	57% $\pm$ 7.420	0.32 ± 0.093	0.57 ± 0.074	0.41 ± 0.087	1	0	56% $\pm$ 14.53	0.32 ± 0.181	0.56 ± 0.145	0.4 ± 0.170	1	0