Robust Watermarking in Singular Values of Contourlet Coefficients

To protect copyright of digital images, many watermarking algorithms have been proposed in recent years. These methods are divided into two groups of spatial and frequency domains. Robustness of a watermark against signal processing modifications is important and most of the existing methods try to address this issue. In this paper a Contourlet based image watermarking algorithm is proposed which embeds a watermark in the singular values of the Contourlet transform of an image. Watermarking is done in three scales of Contourlet transform. Redundant embedding of the watermark in different scales causes higher robustness of the proposed method. In the extraction phase, the final watermark is derived by voting amongst the three extracted watermarks. The experimental results and comparison with the other Contourlet based method shows that the proposed method has better robustness against some of the usual image processing attacks, such as JPEG compression, Salt and Pepper, and Gaussian noise.