yamizi / adversarial-embedding Goto Github PK

Hi,

Was the security of this algorithm only assessed using 154 images and an LSBM detector + SPA?

Why not more images and better steganalysis? e.g. SRM available here.

For more images, I suggest you use the BOSSBase or the ALASKA2 dataset. These are standard datasets used for comparing the security of stego schemes.

Yassine

Hello I am a student in Korea developing watermarking. I reviewed the source at the time and it is very necessary for me. Looking at the code, I have a few questions.

1. How can I deconcode encoded characters?
2. If you look at the paper, the number of classes has been increased by the attack. How can I use the increased number of classes?
3. Would you like to know if you can explain the above description with code?
4. Or what code should be executed to implement the content of the paper?