Table 3.

Performance of proposed detection framework in terms of mAP on GDXray [15] and SIXray [14] datasets using different classification backbones. Bold indicates the best performance.

Network	GDXray [15]	SIXray10 [14]	SIXray100 [14]	SIXray1000 [14]	SIXray [14] *
VGG-16 [48]	0.8691	0.7583	0.5721	0.4126	0.5810
ResNet-50 [46]	0.8917	0.7826	0.6284	0.4392	0.5915
ResNet-101 [46]	0.9162	0.8053	0.6791	0.4527	0.6457

* Average of SIXray10, SIXray100 and SIXray1000 subset.