Reversible Watermarking Using Prediction-Error Expansion and Extreme Learning Machine

Table 1

Comparison between Coltuc’s and proposed schemes for three classical images.

Method	Image	Average improvement (dB)
		Coltuc’s	Proposed
MED	Lena	0.39	0.47
	Jet	0.67	0.41
	Mandrill	0.55	0.61
GAP	Lena	0.15	0.45
	Jet	0.22	0.28
	Mandrill	0.15	0.30