Skip to main content
SpringerLink
Log in

Image tampering detection based on a statistical model

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

Abstract

This paper presents a novel method for image manipulation iden- tification of natural images in JPEG format. The image forgery detection technique is based on a signal-dependent noise model that is relevant to de- scribe a natural image acquired by a digital camera. This parametric model is characterized by two fingerprints which are used to falsification identification. The problem is cast in the framework of the hypothesis testing theory. For practical use, the Generalized Likelihood Ratio Tests (GLRT) are presented and their performance is theoretically established. There are different types of image forgery which have been considered in this paper for example re- sampling, Gaussian filtering and median filtering. Experiments with real and simulated images highlight the relevance of the proposed approach.

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
Fig. 7
Fig. 8

Similar content being viewed by others

References

  1. Bayar B, Stamm MC (2017) On the robustness of constrained convolutional neural networks to jpeg post-compression for image resampling detection. In: acoustics, speech and signal processing (ICASSP), 2017 IEEE international conference on, pp 2152–2156

  2. Doan TNC, Retraint F, Zitzmann C (2017) Blind forensics tool of falsification for im- ages. In: 2017 IEEE international symposium on signal processing and information technology (ISSPIT), pp 018–023

  3. Gloe T, Bohme R (2010) The Dresden image database for benchmarking digital image forensics. Journal of Digital Forensic Practice 3(2–4):150–159

    Article  Google Scholar 

  4. Hwang JJ, Rhee KH (2016) Gaussian filtering detection based on features of residuals in image forensics. In: Computing & Communication Technologies, Research, Innovation, and Vision for the Future (RIVF), 2016 IEEE RIVF International Conference on, pp153–157

  5. Jin J, Jing P, Su Y (2019) AMFNet: an adversarial network for median filtering detection. IEEE Access 6:50459–50467

    Article  Google Scholar 

  6. Julliand T, Nozick V, Talbot H (2016) Automatic image splicing detection based on noise density analysis in raw images. In: International Conference on Advanced Concepts for Intelligent Vision Systems, Springer, pp. 126–134

  7. Le N, Retraint F (2018) Statistical detector of resampled tiff images. In: 2018 IEEE international symposium on signal processing and information technology (ISSPIT), pp 331–335

  8. Le N, Retraint F (2019) An improved algorithm for digital image authentication and forgery localization using Demosaicing artifacts. IEEE Access 7:125038–125053

    Article  Google Scholar 

  9. Liu XL, Lin CC, Yuan SM (2018) Blind dual watermarking for color images au- thentication and copyright protection. IEEE Transactions on Circuits and Systems for Video Technology 28(5):1047–1055

    Article  Google Scholar 

  10. Liu Y, Zhu X, Zhao X, Cao Y (2019) Adversarial learning for constrained image splicing detection and localization based on atrous convolution. IEEE Transactions on Informa- tion Forensics and Security 14(10):2551–2566

    Article  Google Scholar 

  11. Luo S, Peng A, Zeng H, Kang X, Liu L (2019) Deep residual learning using data aug- mentation for median filtering forensics of digital images. IEEE Access 7:80614–80621

    Article  Google Scholar 

  12. Nelder JA, Mead R (1965) A simplex method for function minimization. Comput J 7(4):308–313

    Article  MathSciNet  Google Scholar 

  13. Nguyen HP, Retraint F, Morain-Nicolier F, Delahaies A (2019) An image forgery De- tection solution based on DCT coefficient analysis. International Conference on Information Systems Security and Privacy 2019:487–494

    Article  Google Scholar 

  14. Niu Y, Zhao Y, Ni R (2017) Robust median filtering detection based on local difference descriptor. Signal Process Image Commun 53:65–72

    Article  Google Scholar 

  15. Qiao T, Retraint F (2018) Identifying individual camera device from RAW images. IEEE Access 6:78038–78054

    Article  Google Scholar 

  16. Qiao T, Retraint F, Cogranne R, Thai TH (2017) Individual camera device identification from JPEG images. Sig. Proc.: image Comm. 52: 74–86

  17. Qiao T, Zhu A, Retraint F (2018) Exposing image resampling forgery by using linear parametric model. Multimedia Tools Appl 77(2):1501–1523

    Article  Google Scholar 

  18. Redi JA, Taktak W, Dugelay JL (2011) Digital image forensics: a booklet for beginners. Multimed Tools Appl 51(1):133–162

    Article  Google Scholar 

  19. Rhee KH (2016) Gaussian filtering detection using band pass residual and contrast of forgery image. In: information technology, electronics and Mobile communication conference (IEMCON), 2016 IEEE 7th annual, pp 1–4

  20. Rhee KH (2017) Median filtering detection based on variations and residuals in image forensics. Turkish Journal of Electrical Engineering & Computer Sciences 25(5):3811–3826

    Article  Google Scholar 

  21. Rhee KH (2019) Forensic detection using bit-planes slicing of median filtering image. IEEE Access 7:92586–92597

    Article  Google Scholar 

  22. Romano JP, Lehmann E (2005) Testing statistical hypotheses

  23. Rousseeuw PJ, Croux C (1993) Alternatives to the median absolute deviation. J Am Stat Assoc 88(424):1273–1283

    Article  MathSciNet  Google Scholar 

  24. Scott C (2007) Performance measures for neyman–Pearson classification. IEEE Trans- actions on Information Theory 53(8):2852–2863

    Article  MathSciNet  Google Scholar 

  25. Sheng H, Shen X, Lyu Y, Shi Z, Ma S (2018) Image splicing detection based on Markov features in discrete octonion cosine transform domain. IET Image Process 12(10):1815–1823

    Article  Google Scholar 

  26. Su Y, Jin X, Zhang C, Chen Y (2017) Hierarchical image resampling detection based on blind deconvolution. J Vis Commun Image Represent 48:480–490

    Article  Google Scholar 

  27. Thai TH, Retraint F, Cogranne R (2012) Statistical model of natural images. In: image processing (ICIP), 2012 19th IEEE international conference on, pp 2525–2528

  28. Thai TH, Cogranne R, Retraint F (2014) Camera model identification based on the heteroscedastic noise model. IEEE Trans Image Process 23(1):250–263

    Article  MathSciNet  Google Scholar 

  29. Thai TH, Retraint F, Cogranne R (2015) Generalized signal-dependent noise model and parameter estimation for natural images. Signal Process 114:164–170

    Article  Google Scholar 

  30. Thai TH, Retraint F, Cogranne R (2016) Camera model identification based on the generalized noise model in natural images. Digital Signal Processing 48:285–297

    Article  MathSciNet  Google Scholar 

  31. Yu L, Zhang Y, Han H, Zhang L, Wu F (2019) Robust median filtering forensics by CNN-based multiple residuals learning. IEEE Access 7:120594–120602

    Article  Google Scholar 

Download references

Acknowledgements

This work was supported by ANR project DEFACTO ANR-16-DEFA-0002.

Author information

Authors and Affiliations

  1. University of Technology of Troyes, Troyes, France

    Thi Ngoc Canh Doan &  Florent Retraint

  2. EPF College of Engineering-Montpellier, Montpellier, France

    Cathel Zitzmann

Authors
  1. Thi Ngoc Canh Doan
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Florent Retraint
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Cathel Zitzmann
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Florent Retraint.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

A Proof of Proposition 1

A Proof of Proposition 1

$$ {\displaystyle \begin{array}{l}{V}_{Hj}\left[f\left({\theta}_l;\hat{\tilde{c}}1,{\hat{\tilde{d}}}_1,\Upsilon \right)\right]={V}_{Hj}\left[\frac{1}{\Upsilon^2}{\hat{\tilde{c}}}_1{\theta}_l^{2-2\Upsilon}\right]\\ {}=\frac{\theta_l^{4-2\Upsilon}}{\Upsilon^4}{V}_{Hj}\left[{\hat{\tilde{c}}}_1\right]+\frac{\theta_l^{4-2\Upsilon}}{\Upsilon^4}{V}_{Hj}\left[{\hat{\tilde{d}}}_1\right]\\ {}+\frac{2{\theta}_l^{4-3\Upsilon}}{\Upsilon^4} Co{\upsilon}_{Hj}\left[{\hat{\tilde{c}}}_1,{\hat{\tilde{d}}}_1\right]\\ {}=\frac{\theta_l^{4-2\Upsilon}}{\Upsilon^4}{\upsilon}_{{\tilde{c}}_1}^2+\frac{\theta_l^{4-4\Upsilon}}{\Upsilon^4}{\upsilon}_{{\tilde{d}}_1}^2+2\frac{\theta_l^{4-3\Upsilon}}{\Upsilon^4}{\upsilon}_{{\tilde{c}}_1{\tilde{d}}_1}\end{array}} $$
(22)

Based on the Delta method [18, theorem 11.2.14], we derive

$$ {\displaystyle \begin{array}{c}{V}_{Hj}\left[\mathit{\log}\frac{f\left({\theta}_l;{\tilde{c}}_0,{\tilde{d}}_0,\Upsilon \right)}{f\left({\theta}_l;{\hat{\tilde{c}}}_0,{\hat{\tilde{d}}}_1,\Upsilon \right)}\right]={V}_{Hj}\left[\log \left|f\left({\theta}_l;{\hat{\tilde{c}}}_1,{\hat{\tilde{d}}}_1,\Upsilon \right)\right.\right]\\ {}\approx \frac{V_{Hj}\left[f\left({\theta}_l;{\hat{\tilde{c}}}_1,{\hat{\tilde{d}}}_1,\Upsilon \right)\right]}{f^2\left({\theta}_l;{\tilde{c}}_1,{\tilde{d}}_1,\Upsilon \right)}\end{array}} $$
(23)

and

$$ {\displaystyle \begin{array}{l}{V}_{Hj}\left[\frac{f\left({\theta}_l;{\hat{\tilde{c}}}_0,{\hat{\tilde{d}}}_0,\Upsilon \right)-f\left({\theta}_l;{\tilde{c}}_0,{\tilde{d}}_0,\Upsilon \right)}{f\left({\theta}_l;{\hat{\tilde{c}}}_1,{\hat{\tilde{d}}}_1,\Upsilon \right)f\left({\theta}_l;{\hat{\tilde{c}}}_0,{\hat{\tilde{d}}}_0,\Upsilon \right)}\right]\\ {}={V}_{Hj}\left[\frac{1}{f\left({\theta}_l;{\hat{\tilde{c}}}_1,{\hat{\tilde{d}}}_1,\Upsilon \right)}\right]\approx \frac{V_{Hj}\left[f\left({\theta}_l;{\hat{\tilde{c}}}_0,{\hat{\tilde{d}}}_0,\Upsilon \right)\right]}{f^4\left({\theta}_l;{\tilde{c}}_1,{\tilde{d}}_1,\Upsilon \right)}\end{array}} $$
(24)

Consequently, based on the law of the total variance, it follows that

$$ {\displaystyle \begin{array}{l}{V}_{Hj}\left[\tilde{\tilde{\Lambda}}\left({out}_{l,h}\right)\right]=\frac{1}{2}{\left(\frac{1}{\upsilon_{l,0}^2}-\frac{1}{\upsilon_{l,1}^2}\right)}^2{\upsilon}_{l,j}^4\\ {}+\frac{1}{4}\frac{V_{H_j}\left[f\left({\hat{\theta}}_l;{\hat{\tilde{c}}}_1,{\hat{\tilde{d}}}_1,\Upsilon \right)\right]}{\upsilon_{l,1}^4}\\ {}+\frac{3}{4}\frac{V_{H_j}\left[f\left({\hat{\theta}}_l;{\hat{\tilde{c}}}_1,{\hat{\tilde{d}}}_1,\Upsilon \right)\right]}{\upsilon_{l,1}^8}{\upsilon}_{l,j\cdot}^4\end{array}} $$
(25)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Doan, T.N.C., Retraint, . & Zitzmann, C. Image tampering detection based on a statistical model. Multimed Tools Appl 80, 32905–32924 (2021). https://doi.org/10.1007/s11042-021-11213-3

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11042-021-11213-3

Keywords

Search

Navigation