Table 1. Accuracy in test S-UNIWARD stego-images with different payloads using BOSSBase 1.01 and BOSSBase 1.01 + BOWS 2.

The bold entries indicate the results obtained with the strategy application on the CNN’s. These results show the accuracies on test data for 0.2 bpp and 0.4 bpp.

Dataset	BOSSBase 1.01				BOSSBase 1.01 + BOWS 2
Results	Reported in literature		Strategy		Reported in literature		Strategy
Payload	0.2 bpp	0.4 bpp	0.2 bpp	0.4 bpp	0.2 bpp	0.4 bpp	0.2 bpp	0.4 bpp
Xu-Net	0.6090	0.7280	0.6829	0.7819	–	–	0.7121	0.8182
Ye-Net	0.6000	0.6880	0.7103	0.8101	–	–	0.7269	0.8338
Yedroudj-Net	0.6330	0.7720	0.6773	0.7964	0.6560	–	0.7335	0.8415