Steganalysis via deep residual network

S.T. Wu, S.H. Zhong, Yan Liu

Research output: Chapter in book / Conference proceedingConference article published in proceeding or bookAcademic research


Recent studies have demonstrated that a well designed deep convolutional neural network (CNN) model achieves competitive performances on detecting the presence of secret message in digital images, compared with the classical rich model based steganalysis. In this paper, we propose to investigate a category of very deep CNN model-the deep residual network (DRN), for steganalysis. DRN is suitable for steganalysis from two aspects. For the first, the DRN model usually contains a large number of network layers, which proves to be effective to capture the complex statistics of digital images. For the second, DRN's residual learning (ResL) method actively strengthens the signal coming from secret messages, which is extremely beneficial for the discrimination between cover images and stego images. Comprehensive experiments on standard dataset show that the DRN model achieves very low detection error rates for the state of arts steganographic algorithms. It also outperforms the classical rich model method and several recently proposed CNN based methods.
Original languageEnglish
Title of host publicationProceedings of the International Conference on Parallel and Distributed Systems - ICPADS
Number of pages4
ISBN (Print)9781509044573
Publication statusPublished - 2016
EventInternational Conference on Parallel and Distributed Systems [ICPADS] -
Duration: 1 Jan 2016 → …

Publication series

NameInternational Conference on Parallel and Distributed Systems. Proceedings
ISSN (Print)1521-9097


ConferenceInternational Conference on Parallel and Distributed Systems [ICPADS]
Period1/01/16 → …


  • Steganalysis
  • Convolutional neural network
  • Deep residual network
  • Residual learning
  • Steganography

ASJC Scopus subject areas

  • Hardware and Architecture


Dive into the research topics of 'Steganalysis via deep residual network'. Together they form a unique fingerprint.

Cite this