Table 7.

Performance (in $\%$) on healthy versus non–healthy CXR screening: accuracy, sensitivity, specificity, precision, and F1 score.

Dataset	ACC	AUC	SEN	SPEC	PREC	F1 score
D1	99.87	100.00	99.75	100.00	100.00	99.87
D2	99.55	100.00	99.40	100.00	100.00	99.70
D3	99.76	100.00	99.61	99.91	99.90	99.76
$\mu$	99.72	100.00	99.59	99.97	99.97	99.78
$\sigma$	$\pm$ 0.133	0.00	$\pm$ 0.181	$\pm$ 0.051	$\pm$ 0.058	$\pm$ 0.086