A Digital Image Watermarking Scheme Based on Chaotic Mapping and Singular Value Decomposition

In order to guarantee the security, transparency and robustness of digital image watermark, a scheme based on chaotic mapping and singular value decomposition (SVD) is proposed. Since chaotic mapping is sensitive to initial values; two chaotic maps are employed to scramble the rows and columns of the watermark image, which can enhance the security of the watermark. Initial values of the chaotic mapping are used as private key in the scheme. To utilize the characteristic of the SVD domain for embedding a watermark, the coefficients in the first column of U component are examined. The magnitude difference between the neighboring coefficients is taken as the relationship to embed the watermark. The relationship could be preserved when general image processing is performed. Experimental results demonstrate that the quality of the watermarked image is good and there is strong resistance against general image processing. Furthermore, the scheme can accurately detect and locate the place being tampered in the watermarked image, and the extracted watermark can still be easily identified.