Skip to main content
SpringerLink
Log in

Unsupervised steganalysis over social networks based on multi-reference sub-image sets

  • Published:
Multimedia Tools and Applications Aims and scope Submit manuscript

Abstract

This work proposes a new unsupervised steganalysis scheme which mainly tackles the challenge in identifying individual JPEG image as stego or cover. The proposed scheme does not need a large number of samples to train classification model, and thus it is significantly different from the existing supervised steganalysis schemes. The proposed scheme employs calibration technology to construct multiple reference images from one suspicious image. These reference images are considered as the imitation of cover. Furthermore, randomized sampling is performed to construct sub-image sets from suspicious image and reference images, respectively. By calculating the maximum mean discrepancy between any two sub-image sets, an efficient measure is provided to give the optimal decision on this suspicious image. Experimental results show that the proposed scheme is effective and efficient in identifying individual image, and outperforms the state-of-the-art steganalysis scheme.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6

Similar content being viewed by others

Notes

  1. We assume that the images from the same camera have the same noise distribution.

  2. A fixed cropped size is used to show the overall performance of proposed scheme, although there might exist an optimal range for the cropped size m×n for the recent image database.

  3. In fact, D value can be changed in smaller steps, e.g. 0.01 step.

  4. Since the scale of training set can potentially affect the matching degree of generalized classification model in supervised schemes and further affect the accuracy of classification, we tend to use the same number of images as the training set to show a fair comparison.

References

  1. Böhme R (2007) Weighted stego-image steganalysis for JPEG covers. 10th International Workshop on information hiding, Barbara, USA, May 19-21, 5284:178-194

  2. Chang C, Lin C (2001) LIBSVM: a library for support vector machines 2001. Software available at http://www.csie.ntu.edu.tw/~cjlin/libsvm

  3. Cogranne R, Zitzmann C, Retraint F, Nikiforov I, Cornu P, Fillatre L (2014) A local adaptive model of natural images for almost optimal detection of hidden data. Elsevier Signal Process 100:169–185

    Article  Google Scholar 

  4. Denemark T, Fridrich J, Holub V (2014) Further study on security of S-UNIWARD. Proceedings of SPIE, Electronic Imaging, Media Watermarking, Security, and Forensics, 9028, San Francisco, CA, February 2-6

  5. Fridrich J, Kodovský J (2012) Rich models for steganalysis of digital images. IEEE Trans Inf Forensics Secur 7(3):868–882

    Article  Google Scholar 

  6. Fridrich J, Goljan M, Hogea D (2002) Steganalysis of JPEG images: breaking the F5 algorithm, 5th International Workshop on information hiding, Noordwijkerout, Netherlands, October 7-9, 2578:310–323

  7. Fridrich J, Pevný T, Kodovský J (2007) Statistically undetectable JPEG steganography: dead ends, challenges, and opportunities. Proceedings of the 9th ACM multimedia \& security Workshop, Dallas, TX, September 20-21, pp 3–14

  8. Gretton A, Borgwardt K, Rasch M, Scholkopf B, Smola A (2007) A kernel method for the two-sample-problem. Adv Neural Inf Proces Syst 19, MIT press, Cambridge, 513-520

  9. Hetzl S, Mutzel P (2005) A graph-theoretic approach to steganography. Proceedings of 9th IFIP TC-6 TC-11 Internetional conference on communications and multimedia security, Salzburg, Austria, September 19-21, 3677:119-128

  10. Holub V, Fridrich J (2013) Random projections of residuals for digital image steganalysis. IEEE Trans Inf Forensics Secur 8(12):1996–2006

    Article  Google Scholar 

  11. JSteg algorithm. Available from: http://zooid.org/%7epaul/crypto/jsteg/. Accessed on 2011

  12. Ker A, Pevný T (2011) A new paradigm for steganalysis via clustering. Proc SPIE, Media Watermarking, Secur, Forensics III 7880:78800U

    Article  Google Scholar 

  13. Ker A, Pevný T (2012) Identifying a steganographer in realistic and hereogeneous data sets. Proc SPIE, Media Watermarking, Secur, Forensics XIV 8303:N01–N13

    Google Scholar 

  14. Ker A, Pevný T (2014) The steganographer is the outlier: realistic large-scale steganalysis. IEEE Trans Inf Forensics Secur 9(9):1424–1435

    Article  Google Scholar 

  15. Kodovský J, Fridrich J (2009) Calibration revisited. Proceedings of the 11th ACM multimedia \& security Workshop, 63-74

  16. Kodovský J, Fridrich J (2010) Quantitative structural steganalysis of Jsteg. IEEE Trans Inf Forensics Secur 5(4):681–693

    Article  Google Scholar 

  17. Li F, Zhang X, Cheng H, Yu J (2016) Digital image steganalysis based on local textural features and double dimensionality reduction. Secur Commun Netw 9(8):729–736

    Article  Google Scholar 

  18. Li F, Wu K, Lei J, Wen M, Bi Z, Gu C (2016) Steganalysis over large-scale social networks with high-order joint features and clustering ensembles. IEEE Trans Inf Forensics Secur 11(2):344–357

    Article  Google Scholar 

  19. Li F, Wu K, Lei J, Wen M, Cao Y, Jin T (2016) Multi-source stego detection with local textural feature and clustering ensembles. Technical Report, Sept., 2016

  20. [Online]. http://www.instagram.com. Accessed on March 2016

  21. [Online]. Available: http://exile.felk.cvut.cz/boss/BOSSFinal/

  22. [Online]. Available: http://dde.binghamton.edu/download/

  23. [Online]. http://www.flickr.com. Accessed on March 2016

  24. Pevný T, Fridrich J (2007) Merging markov and DCT features for multi-class JPEG Steganalysis. Proceedings of SPIE, Electronic imaging, security, steganography, and Watermarking of multimedia contents IX, San Jose, CA, January 29- February 1, 6505: 3 1-3 14

  25. Pevný T, Bas P, Fridrich J (2010) Steganalysis by subtractive pixel adjacency matrix. IEEE Trans Inf Forensics Secur 5(2):215–224

    Article  Google Scholar 

  26. Qiao T, Retraint F, Cogranne R, Zitzmann C (2015) Steganalysis of JSteg algorithm using hypothesis testing theory. EURASIP J Inf Secur 2(1):1–16

    Google Scholar 

  27. Qin C, Hu Y (2016) Reversible data hiding in VQ index table with lossless coding and adaptive switching mechanism. Signal Process 129:48–55

    Article  Google Scholar 

  28. Shi Y, Chen C, Chen W (2006) A markov process based approach to effective attacking JPEG steganography, 8th International Workshop on information hiding, Alexandria, VA, July 10-12, 4437: 249–264

  29. Thai T, Cogranne R, Retraint F (2014) Statistical model of quantized DCT coefficients: application in the steganalysis of Jsteg algorithm. IEEE Trans Image Process 23(5):1980–1993

    Article  MathSciNet  MATH  Google Scholar 

  30. Westfeld A (2001) High capacity despite better steganalysis (F5-a steganographic algorithm). Proceedings of 4th International Workshop on information hiding. Pittsburgh, PA, April 25-27, pp 289–302

  31. Xia Z, Wang X, Sun X, Wang B (2014) Steganalysis of least significant bit matching using multi-order differences. Secur Commun Netw 7(8):1283–1291

    Article  Google Scholar 

  32. Xia Z, Wang X, Sun X, Liu Q, Xiong N (2016) Steganalysis of LSB matching using differences between nonadjacent pixels. Multimed Tools Appl 75(4):1947–1962

    Article  Google Scholar 

  33. Zhang T, Ping X (2003) A fast and effective steganalytic technique against Jsteg-like algorithm. Proceedings of the ACM Symposium on Applied Computing, New York, USA, 307–311

Download references

Acknowledgments

This work was supported by Natural Science Foundation of China under Grants (No. 61602295, 61373152, 61672337, 61572311, 61572309, and 61472236), Natural Science Foundation of Shanghai (No. 16ZR1413100), the Scientific Research Foundation of Shanghai University of Electric Power (No. K2015-010), the Excellent University Young Teachers Training Program of Shanghai Municipal Education Commission (No. ZZsdl15105), Project of Shanghai Science and Technology Committee (14110500800) and the “Dawn” Program of Shanghai Education Commission (No. 16SG47).

Author information

Authors and Affiliations

  1. College of Computer Science and Technology, Shanghai University of Electric Power, Shanghai, 200090, People’s Republic of China

    Fengyong Li, Kui Wu, Jingsheng Lei & Mi Wen

  2. Department of Computer Science, University of Victoria, Victoria, B.C., Canada

    Kui Wu

  3. School of Communication and Information Engineering, Shanghai University, Shanghai, 200444, People’s Republic of China

    Yanli Ren

Authors
  1. Fengyong Li
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Kui Wu
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Jingsheng Lei
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Mi Wen
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Yanli Ren
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Fengyong Li.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Li, F., Wu, K., Lei, J. et al. Unsupervised steganalysis over social networks based on multi-reference sub-image sets. Multimed Tools Appl 77, 17953–17971 (2018). https://doi.org/10.1007/s11042-017-4759-x

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11042-017-4759-x

Keywords

Search

Navigation