Adaptive and Perceptual Watermarking of Still Images. Tatouage Perceptuel et Adaptatif D’images Fixes

Adaptive and Perceptual Watermarking of Still Images

Tatouage Perceptuel et Adaptatif D’images Fixes

A. Saadane F. Autrusseau 

Institut de Recherche en Communications et Cybernétique de Nantes (IRCCyN), La Chantrerie,BP50609, 44306 Nantes Cedex

Page: 
235-247
|
Received: 
N/A
|
Accepted: 
N/A
|
Published: 
31 August 2001
| Citation

OPEN ACCESS

Abstract: 

This paper presents a new adaptive and perceptual watermarking algorithm. This algorithm is called perceptual as it uses a model of the human visual system (HVS) to determine the auspicious sites for watermarking. The HVS modelisation considered here is consistent with a decomposition that uses a non directional low pass channel and a set of three band pass radial frequency channels each being decomposed into angular sectors. The watermarking is also called adaptive as it exploits an error visibility model to compute for each image and for each selected site the maximum watermark strength to be applied without inducing visible degradations. The algorithm performances have been evaluated in terms of watermark invisibility and robustness to different attacks. In the first case, subjective tests, based on CCIR recommandation, have been conducted to assess visual quality of images watermarked with different strengths. In the second case, the correlation coefficient is used to determine the original watermark detection efficiency to attacks such as filtering, noise addition, JPEG compression, pseudo-cropping and limited geométric distorsions.

Résumé

Cet article présente un nouvel algorithme de tatouage perceptuel et adaptatif. Il est perceptuel parce qu’il exploite une modélisation du comportement du système visuel humain pour déterminer les sites propices au tatouage. La modélisation considérée ici, décompose l’espace de représentation en 17 canaux visuels. Ces derniers se répartissent en un canal basses fréquences non directionnel et trois bandes de fréquences radiales, elles mêmes décomposées en canaux angulaires dont le nombre dépend de la bande radiale considérée. Le tatouage est dit également adaptatif parce qu’il utilise un modèle de perception des erreurs pour calculer la force maximale à appliquer pour que l’intégration du filigrane n’engendre pas de dégradations visibles. Les performances de cette approche ont été évaluées en termes de transparence du filigrane et de sa robustesse à différentes attaques. Dans le premier cas, des tests d’évaluation menés selon la recommandation 500 du CCIR ont permis de juger en fonction de la force du filigrane, la qualité visuelle des images tatouées par rapport à celle des images originales. Dans le deuxième cas, le calcul du coefficient de corrélation a permis d’analyser l’efficacité du recouvrement du filigrane original face à des attaques comme le filtrage passe-bas, le filtrage median, l’ajout de bruit, la compression JPEG à différents taux, le pseudo-cropping et les attaques géométriques limitées.

Keywords: 

Perceptual watermarking, Human Visual System, Error Visibility model, masking effects.

Mots clés 

tatouage perceptuel, Système Visuel Humain, modèle de visibilité des erreurs, effets de masquage.

1. Introduction
2. Modélisation du SVH
3. Approche Proposée
4. Résultats
5. Conclusion
6. Remerciements
  References

[1] R. Schyndel, A.Z. Tirkel and C.F. Osborne, « A digital watermark », Proceedings of the International Conference on Image Processing 2, pp. 86-90, 1994.

[2] I.J. Cox, J. Kilian, T. Leighton and T. Shamoon, « Secure spread spectrum watermarking for images, audio and video », Proceedings of the International Conference on Image Processing, pp. 243-246, 1996. 

[3] D. Kundur and D. Hatzinakos, « Digital watermarking using miultiresolution wavelet decomposition », Proceedings of the ICASSP 5, pp. 2659-2662, Mai 1998. 

[4] M. Kutter, « Watermarking resisting to translation, rotation, and scaling », Proceedings of SPIE, Multimedia Systems and Applications 3528, pp. 423-431, November 1998. 

[5] G. Langelaar, R. Langendijk and J. Biemond, « Watermarking by DCT coefficient removal : Statistical approach to optimal parameter settings », Proc. SPIE IST/SPIE’s 11th Annu., Symp., Electronic Imaging : Security and Watermarking of multimedia contents, Vol. 3657, 1999. 

[6] I.N. Nikolaidis, « Copyright protection of images using robust digital signatures », Proceedings of ICASSP, 1996. 

[7] M. Barni, F. Bartolini, V. Cappellini, and A. Piva, « A DCT-domain system for robust image watermarking », Signal Processing 66, pp. 357-372, 1998. 

[8] K. Matsui, K. Tanaka and Y. Nakamura, « Digital signature on a fascimile document by recursive MH coding », Sympos. on cryptography and information security, 1989. 

[9] K. Tanaka, Y. Nakamura and K. Matsui, « Embedding secret information into a dithered multi-level image », Proc. IEEE Military Communications Conf., pp. 216-220, 1990. 

[10] J. Ruanaidh, F.M. Boland and O. Sinnen, « Watermarking digital images for copyright protection », Proc Electronic Imaging and the Visual Arts, 1996. 

[11] A.piva, M. Barni, F. Bartiolini and V. Cappellini, « Threshold selection for correlation-based watermark detection », Proceedings COST 254 Workshop on Intelligent Communications, pp. 67-72, June 4-6 1998. 

[12] D. Benham, N. Memon and M.B.-L. Yeo, « Fast waremarking of DCTbased compressed images », Proc. Int. Conf. Image Science, Systems, and Technology, pp. 243-253, 1997. 

[13] F. Hartung and B. Girod, « Digital watermarking of raw and compressed video », Proc. SPIE compression technologies and systems for video comm., Vol. 2952, pp. 205-213, 1996. 

[14] E. Koch and J. Zhao, « Toward robust and hidden image copyright labeling », Proc. Workshop nonlinear signal and image processing, 1995. 

[15] G.C. Langelaar, J.C.A. V. der Lubbe and R.C. Lagendijk, « Robust labeling method for copy protection of images », Storage and retrieval for Image and Video databases V, 3022, pp. 298-309, 1997. 

[16] D. Kundur and D. Hatzinakos, « A robust digital image watermarking method uding wavelet-based fusion », Proceedings of the International Conference on Image Processing 1, pp. 544-547, 1997. 

[17] H.-J. Wang and C.-C. J. Kuo, « An integrated progressive image coding and watermark system », Proceedings of the ICASSP, Vol. 5, pp. 3721-3724, 1998. 

[18] P. Bas, J. Chassery and F. Davoine, « Using the fractal code to watermark images », Proceedings of the International Conference on Image Processing 1, pp. 470-474, 1998. 

[19] P. Prandoni and M. Vetterli, « Perceptually hidden data transmission over audio signals », IEEE ICASSP, Vol. 6, pp. 2665-3668, 1998. 

[20] J. Lacy, S. Quackenbush, A.R. Reibman, D. Shur and J. Suyder, « On combining watermarking with perceptual coding », Proceedings of ICASSP, Vol. 6, pp. 3725-3728, 1998. 

[21] J.R. Hernadez, F. Perez-Gonzalez and J.M. Rodriguez, « The impact of chanel coding on the performance of spatial watermarking for copyright protection », Proceedings of ICASSP, Vol. 5, pp. 2973-2976, 1998. 

[22] J.F. Delaigle, C.D. Vleeschouwer and B. Macq, » Watermarking algorithm based on a human visual model », Signal Processing, Vol. 66, pp. 319-335, 1998. 

[23] M. Swanson, B. Zhu, A.H. Tewfik and L. Boney, « Robuste audio watermarking using perceptual coding », Signal Processing (Special issue on watermarking), Vol. 66, pp. 337-356, 1998. 

[24] I. Pitas, « A method for signature casting on digital images », Proceedings of ICIP, IEEE press, Vol. 3, pp. 215-218, 1996. 

[25] J.L. Mannos and D.J. Sakrison, « The effects of a visual fidelity criterion on the encoding of images », IEEE transactions on Information Theory, Vol. IT-20, pp. 525-536, July 1974.

[26] P. Barten, « Evaluation of subjective image quality with the square-root integral method », Journal of Optical Society of America 7, pp. 20242031, 1990. 

[27] A. De Rosa, M. Barni, F. Bartolini and A. Piva, « Watermark capacity measure incorporating in a model of the human visual system », IST/SPIE’s 13th International Symposium Electronic Imaging : Multimedia Processing and applications (Security and watermarking of multimedia contents III), 2001. 

[28] D.J. Sakrison, « On the role of the observer and a distorsion measure in image transmission », IEEE trans. on Com., Vol. 25, No 11, pp. 1251-1267, 1977. 

[29] J.N. Graham, « Detection of gratting patterns containing two spatial frequen,cies ; A comparison of a single channel and multiple channels models », Vis. Research, Vol. 11, pp. 251-259, 1971.

[30] R. Valois and K. Valois, « Spatial vision », Oxford Univ. Press, 1988. 

[31] C.D. Burr and S.A. Wiijensundra, « Orientation discrimination depends on spatial frequency », Vis. Res., Vol. 31, No 7/8, pp. 1449-1452, 1991. 

[32] S.J. Anderson, Burr D.C. and M. Morrone, « Two dimensional spatial frequency selectivity of motion sensitive mechanisms in human vision », J.O.S.A., Vol. 8, pp. 1340-1351, 1991. 

[33] G.C. Philips and H. Wilson, « Orientation bandwidths of spatial mechanisms measured by masking », J.O.S.A., Vol. 1, No 2, pp. 226-232, 1984. 

[34] S. Daly, « The visible difference predictor : An algorithm for the assessment of image fidelity », Proc. of SPIE, Human Vision, Visual Processing and Digital Display, Vol. III, pp. 2-15, 1992. 

[35] A. Watson, « The cortex transform : Rapid computation of simulated neural images », Computer Vision and Image Processing, 39, pp. 311-327, 1987. 

[36] M.A. Georgeson and M. Harris, « Spatial selectivity of contrast adaptation : Models and data », Vision Research, Vol. 24, pp. 729-749, 1984. 

[37] J.G. Daugman, « Spatial visual channels in the fourier plane », Vis. Research, Vol. 24, No 9, pp. 891-910, 1984. 

[38] A. Saadane, D. Barba and H. Senane, « The estimation of visual bandwidthsand their impact in image decomposition and coding », Proceedings of Visual Communications and Image Processing, 1993. 

[39] G.E. Legge and J.M. Foley, « Contrast masking in Human Vision », Journal of the Optical Soc. of America, 70 (No 12), pp. 1458-1471, 1980. 

[40] J.M. Foley, « Human luminance pattern mechanisms : Masking experiments require a new model », J.O.S.A. A 11 (6), pp. 171-1719, 1994. 

[41] D.J. Heeger, « Normalisation of cells responses in cat striate cortex », Visual Neuroscience, Vol. 9, pp. 181-198, 1992. 

[42] P.C. Theo and D.J. Heeger, « Perceptual image distorsion » Proc. of SPIE, Vol. 2179, pp. 127-141, 1994. 

[43] H.R. Wilson and J.R. Bergen, « A four mechanism model for threshold spatial vision », Vis. Res., pp. 19-32, 1979. 

[44] H.R. Wilson D. McFarlane and G.C. Philips, « Spatial frequency tuning of orientation selective inuts estimated by oblique masking », Vis. Res.,Vol. 23, pp. 873-82, 1983. 

[45] A. Saadane, N. Bekkat and D. Barba, « On the masking effects in a perceptually based image quality metric », Advances in the theory of computation and computational mathematics book serties, Vol. Imaging and Vision Systems, 2001. 

[46] CCIR, « Projet de revision de la recommandation 500-4 : méthode d’évaluation subjective de la qualité des images de télévision », Document commissions d’études du CCIR 11/BL/51-F, 1992.

[47] C.I. Poldichuck and W. Zeng, « Image-adaptive watermarking using visual models », IEEE Journal on Selected Areas in Communications, 16, pp. 525-539, May 1998. 

[48] G.E. Legge, « Spatial frequency masking in human vision : Binocular interactions », J. OPt. Soc. Amer., A 69 (6), pp. 838-847, 1979. 

[49] A.B. Watson, « DCT quantization matrices visually optimized for individual images », Human Vision, Visual Processing and Digital Display IV, Proc. SPIE, Vol. 1913, pp. 202-216, 1993.