Multimedia Tools and Applications

, Volume 78, Issue 10, pp 13805–13818 | Cite as

Generative steganography with Kerckhoffs’ principle

  • Yan KeEmail author
  • Min-qing Zhang
  • Jia Liu
  • Ting-ting Su
  • Xiao-yuan Yang


The distortion in steganography that usually comes from the modification or recoding of the cover image during the embedding process. And it is the embedding distortion that leaves the steganalyzer with possible discrimination. Therefore, we propose generative steganography with Kerckhoffs’ principle (GSK) in this paper. In GSK, the secret messages are generated by a cover image using a generator rather than embedded into the cover, which results in no modifications to the cover. To ensure security, the generators are trained to meet Kerckhoffs’ principle based on generative adversarial networks (GANs). Everything about the GSK system is public knowledge for the receivers, except the extraction key. The secret messages can be outputted by the generator if and only if the extraction key and the cover image are both inputted. In the generator training procedures, there are two GANs (Message-GAN and Cover-GAN) that are designed to work jointly, making the generated results under the control of the extraction key and the cover image. We provide experimental results for the training process. We present an example of the working process by adopting a generator trained on the dataset MNIST, which demonstrates that GSK can use a cover image without any modification to generate messages. Furthermore, only meaningless results would be obtained without the extraction key or the cover image.


Information security Generative steganography Generative adversarial networks (GANs) Kerckhoffs’ principle 



This work was supported in part by the National Key R&D Program of China under Grant 2017YFB0802000.


  1. 1.
    Abadi M et al (2016) Learning to protect communications with adversarial neural cryptography. arXiv: 1610.06918v1Google Scholar
  2. 2.
    Cayre F, Bas P (2008) Kerckhoffs-based embedding security classes for woa data hiding. IEEE Transactions on Information Forensics and Security 3(1):1–15CrossRefGoogle Scholar
  3. 3.
    Cayre F, Fontaine C, Furon T (2005) Watermarking security: theory and practice. IEEE Trans Signal Process 53(10):3976–3987MathSciNetCrossRefzbMATHGoogle Scholar
  4. 4.
    Chang C-C, Nguyen T-S, Chien T-Y (2017) An efficient steganography scheme based on edge detection for high payload. Journal of Information Hiding and Multimedia Signal Processing 8(5):967–979Google Scholar
  5. 5.
    Chen X et al (2016) InfoGAN: Interpretable representation learning by information maximizing generative adversarial nets. arXiv: 1606.03657Google Scholar
  6. 6.
    Fang D, Sun S (2018) A new scheme for image steganography based on Hyperchaotic map and DNA sequence. Journal of Information Hiding and Multimedia Signal Processing 9(2):392–399Google Scholar
  7. 7.
    Filler T, Judas J, Fridrich J (2011) Minimizing additive distortion in steganography using syndrome-trellis codes. IEEE Trans Inf Forensics Security 6(3):920–935CrossRefGoogle Scholar
  8. 8.
    Furon T (2002) Security analysis, European Project IST-1999-10987 CERTIMARK, Deliverable D.5.5Google Scholar
  9. 9.
    Goodfellow I et al (2014) Generative adversarial nets. In: Proc. Adv. Neural Inf. Process. Syst., pp 2672–2680Google Scholar
  10. 10.
    Kerckhoffs A (1883) La cryptographie militaire. Journal des sciences militaires IX:5–83. pp 161–191, February 1883Google Scholar
  11. 11.
    Li B, He J, Huang J, Shi Y (2011) A survey on image steganography and Steganalysis. Journal of information Hiding and Multimedia Signal Processing 2(2):142–172Google Scholar
  12. 12.
    Mirza M, Osindero S (2014) Conditional generative adversarial nets. arXiv preprint arXiv: 1411.1784Google Scholar
  13. 13.
    Odena A et al (2017) Conditional image synthesis with auxiliary classifier GANS. arXiv: 1610.09585v3Google Scholar
  14. 14.
    Ou B, Li X, Zhao Y, Ni R, Shi Y-Q (2013) Pairwise prediction-error expansion for efficient reversible data hiding. IEEE Trans Image Process 22(12):5010–5021MathSciNetCrossRefzbMATHGoogle Scholar
  15. 15.
    Qian Y et al (2016) Learning and transferring representations for image steganalysis using convolutional neural network. In: Proc. IEEE 2016 Int. Conf. Image Process, pp 2752–2756Google Scholar
  16. 16.
    Radford A, Metz L, Chintala S (2015) Unsupervised representation learning with deep convolutional generative adversarial networks. arXiv: 1511.06434Google Scholar
  17. 17.
    Swire PP (2004) A model for when disclosure helps security: what is different about computer and network security? Journal on Telecommunications and High Technology Law 1(3):163–208Google Scholar
  18. 18.
    Xu G, Wu HZ, Shi YQ (2016) Structural design of convolutional neural networks for steganalysis. IEEE Signal Process Lett 23(5):708–712CrossRefGoogle Scholar
  19. 19.
    Xue P-F, Hu J-S, Hu R-G, Liu H-L, Yu G (2018) A new DHT: network steganography based on distributed coding. Journal of Information Hiding and Multimedia Signal Processing 9(2):355–369Google Scholar

Copyright information

© Springer Science+Business Media, LLC, part of Springer Nature 2018

Authors and Affiliations

  • Yan Ke
    • 1
    Email author
  • Min-qing Zhang
    • 1
  • Jia Liu
    • 1
  • Ting-ting Su
    • 1
  • Xiao-yuan Yang
    • 1
  1. 1.Key Laboratory of Network and Information Security Under the Chinese People Armed Police Force (PAP)Engineering University of PAPXi’anChina

Personalised recommendations