[Retracted] Detecting Digital Watermarking Image Attacks Using a Convolution Neural Network Approach

Figure 12

Extract attack from watermark images with a size of 64 × 64; (a) median filter; (b) Gaussian filter; (c) salt-and-pepper; (d) average filter (e) motion blur; (f) no attack.

(a)

(b)

(c)

(d)

(e)

(f)