Research Article
Zero-Watermarking Algorithm for Medical Image Based on VGG19 Deep Convolution Neural Network
Table 5
Experimental results under spherical distortion attack.
| ā | Distortion quantity (%) | 10 | 30 | 50 | 70 | 90 |
| Medical image A | PSNR (dB) | 12.8585 | 11.4917 | 11.1459 | 11.3035 | 11.3435 | NC | 1.0 | 1.0 | 0.87166 | 0.87166 | 0.87166 |
| Medical image B | PSNR (dB) | 18.4985 | 14.4427 | 13.0098 | 11.7524 | 10.5874 | NC | 0.93654 | 0.87366 | 0.81221 | 0.81221 | 0.74704 |
| Medical image C | PSNR (dB) | 27.259 | 19.872 | 16.6028 | 14.5081 | 13.0453 | NC | 0.93512 | 1.0 | 1.0 | 1.0 | 0.93997 |
| Medical image D | PSNR (dB) | 23.9151 | 17.1131 | 14.5852 | 13.2542 | 12.4597 | NC | 1 | 1 | 1 | 1 | 0.93769 |
| Medical image E | PSNR (dB) | 30.64 | 25.4487 | 22.91 | 20.9156 | 19.2908 | NC | 0.9394 | 1 | 1 | 0.81763 | 0.62984 |
|
|