Steganography of Steganographic Networks

This repo is the official code for

• Steganography of Steganographic Networks.

• Python 3.8.13, PyTorch = 1.11.0

• Run the following commands in your terminal:

  conda env create -f env.yaml

  conda activate pyt_env

• Run python train_secret_model.py,

• Run python secret_model_disguising.py,

• Run python secret_model_extraction.py,

1. The secret and stego-models will be saved in 'checkpoint/'
2. The results and running logs will be saved in 'results/'

• VGG11 is a simplified version that employs VGG11 as the backbone, excluding batch normalization layers and skip connections (shortcuts). To understand our code, we recommend you starting with VGG11.

• We are organizing the code about "HiDDeN" and plan to upload it in the near future.

If you find our paper or code useful for your research, please cite:

@inproceedings{li2023steganography,
  title={Steganography of steganographic networks},
  author={Li, Guobiao and Li, Sheng and Li, Meiling and Zhang, Xinpeng and Qian, Zhenxing},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  volume={37},
  number={4},
  pages={5178--5186},
  year={2023}
}