Zero-Watermarking Algorithm for Medical Image Based on VGG19 Deep Convolution Neural Network

Table 3

Experimental results under extrusion distortion attack.

	Distortion quantity (%)	10	30	50	70	90
Medical image A	PSNR (dB)	13.0494	11.4873	10.6621	10.2615	9.9121
	NC	1.0	1.0	1.0	0.9354	0.9354
Medical image B	PSNR (dB)	21.7239	16.4862	14.1855	12.7081	11.6212
	NC	1.0	0.84222	0.7815	0.81049	0.81278
Medical image C	PSNR (dB)	26.561	19.8342	17.1758	15.695	14.9073
	NC	0.93826	0.93826	0.74818	0.74818	0.81163
Medical image D	PSNR (dB)	24.4511	18.0799	15.8822	14.692	14.0065
	NC	0.93626	0.81021	0.78076	0.78076	0.87395
Medical image E	PSNR (dB)	32.0185	26.229	23.5813	21.8296	20.504
	NC	0.9394	0.9394	0.9394	0.87566	0.87766