Table 2.

Classification performance for models of configuration S+No-FT, using the VGG-19 architecture

Metric	INbreast models	CBIS-DDSM models
	$\overline{x}$	s	$\overline{x}$	s
G-Mean	0.3773	0.1043	0.3476	0.2534
F2-Score	0.1882	0.0625	0.1347	0.1148
Accuracy	0.2183	0.0602	0.7379	0.0678
Recall	0.7667	0.2509	0.2333	0.1876
Specificity	0.1901	0.0558	0.7639	0.0707
Precision	0.0470	0.0160	0.0517	0.0467