Research Article
[Retracted] Detecting Digital Watermarking Image Attacks Using a Convolution Neural Network Approach
Table 3
Results of the CNN model attack with a watermark image size of 256 × 256.
| Attacks | PSNR | SSIM |
| Median filter | 36.32 | 98.62 | Gaussian filter | 35.72 | 98.17 | Salt-and-pepper | 33.50 | 97.11 | Average filter | 35.66 | 98.13 | Motion blur | 32.35 | 97.09 | No attack | 37.61 | 99.99 |
|
|