This is the official code for 《Individualized Image Steganography Method with Dynamic Separable Key and Adaptive Redundancy Anchor》

Before you start training, prepare Paris StreetView dataset and ImageNet dataset, or any datasets you want.

Then, change the content in config.py to set everything matching your environment.

Now you can start by running "python train_x.py" where x represents the number of secret images. We prepare codes for x = 1, 2 and 3. You can simply edit code hidng more secret images.