Table 4.

Fivefold cross-validation results of the second scenario (pretrained X-ray models) using three CNN architectures—ResneXt-50, Densenet-161, and Inception-v3—and two loss functions (MSE, Huber). ^§ Mean and STD were calculated using the average and the standard deviation of the five folds results, respectively. ^* the evaluation metrics were calculated using all five folds predictions. The bold is for the results of the five folds. The red, purple and blue colors are for the best performances for Mean (5 folds), STD (5 folds) and All Predictions, respectively.

Architecture	Fold	PC ↑	MAE ↓	RMSE ↓	${\mathit{P}\mathit{C}}_{\mathit{s}\mathit{u}\mathit{b}\mathit{j}}$↑	${\mathit{M}\mathit{A}\mathit{E}}_{\mathit{s}\mathit{u}\mathit{b}\mathit{j}}$↓	${\mathit{R}\mathit{M}\mathit{S}\mathit{E}}_{\mathit{s}\mathit{u}\mathit{b}\mathit{j}}$↓
	Fold 1	0.9453	5.24	8.77	0.9619	4.05	6.58
	Fold 2	0.8569	6.68	13.39	0.9234	4.87	8.70
ResneXt-50	Fold 3	0.9083	6.17	12.53	0.9196	5.23	11.17
MSE	Fold 4	0.9589	3.97	6.32	0.9884	2.31	3.22
	Fold 5	0.9457	5.99	9.05	0.9674	4.92	6.43
	Mean (5 folds) §	0.9230	5.61	10.01	0.9521	4.28	7.22
	STD (5 folds) §	0.0371	0.9412	2.6019	0.0265	1.0577	2.6391
	All Predictions *	0.9173	5.62	10.41	0.9431	4.29	7.91
	Fold 1	0.9487	4.96	8.48	0.9633	3.80	6.43
	Fold 2	0.8227	6.93	15.29	0.9147	4.96	9.25
ResneXt-50	Fold 3	0.9191	5.35	11.84	0.9301	4.27	10.47
Huber	Fold 4	0.9622	3.70	6.09	0.9873	2.39	3.26
	Fold 5	0.9507	5.50	8.80	0.9703	4.34	6.30
	Mean (5 folds) §	0.9207	5.29	10.10	0.9532	3.95	7.14
	STD (5 folds) §	0.0510	1.0367	3.1735	0.0267	0.8637	2.5221
	All Predictions *	0.9132	5.30	10.65	0.9427	4.02	7.87
	Fold 1	0.9438	5.15	8.65	0.9611	3.79	6.36
	Fold 2	0.8768	6.48	12.34	0.9354	4.62	8.04
Densenet-161	Fold 3	0.9116	5.86	12.44	0.9259	4.73	10.89
MSE	Fold 4	0.9576	4.11	6.65	0.9861	2.97	4.14
	Fold 5	0.9479	5.79	8.99	0.9691	4.82	6.38
	Mean (5 folds) §	0.9275	5.48	9.81	0.9555	4.18	7.16
	STD (5 folds) §	0.0297	0.8032	2.2503	0.0220	0.7101	2.2385
	All Predictions *	0.9234	5.49	10.12	0.9501	4.09	7.56
	Fold 1	0.9440	4.97	8.78	0.9569	3.91	6.84
	Fold 2	0.8938	6.14	11.36	0.9459	4.57	7.58
Densenet-161	Fold 3	0.9158	5.89	12.33	0.9290	4.99	10.87
Huber	Fold 4	0.9643	3.67	5.93	0.9893	2.33	3.16
	Fold 5	0.9526	5.48	8.71	0.9697	4.56	6.54
	Mean (5 folds) §	0.9341	5.23	9.42	0.9582	4.07	7.00
	STD (5 folds) §	0.0257	0.8749	2.2505	0.0205	0.9369	2.4615
	All Predictions *	0.9305	5.25	9.75	0.9523	4.06	7.51
	Fold 1	0.9483	5.22	8.51	0.9630	3.95	6.53
	Fold 2	0.8571	6.63	13.14	0.9258	5.01	8.57
Inception-v3	Fold 3	0.9111	5.89	12.28	0.9240	5.22	10.84
MSE	Fold 4	0.9605	3.91	6.22	0.9878	2.52	3.27
	Fold 5	0.9463	6.10	9.22	0.9667	5.02	6.74
	Mean (5 folds) §	0.9247	5.55	9.87	0.9535	4.34	7.19
	STD (5 folds) §	0.0375	0.9362	2.5335	0.0248	1.0149	2.4992
	All predictions	0.9163	5.90	10.57	0.9470	4.16	7.51
	Fold 1	0.9526	4.81	8.43	0.9673	3.81	6.45
	Fold 2	0.8905	6.17	11.64	0.9414	4.59	7.63
Inception-v3	Fold 3	0.9230	5.47	11.69	0.9324	4.75	10.41
Huber	Fold 4	0.9607	3.79	6.33	0.9850	2.85	3.91
	Fold 5	0.9556	5.29	8.14	0.9753	4.04	5.56
	Mean (5 folds) §	0.9365	5.10	9.25	0.9603	4.01	6.79
	STD (5 folds) §	0.0264	0.7896	2.1022	0.0201	0.6738	2.1786
	All Predictions *	0.9330	5.34	9.44	0.9558	3.98	6.87