Robust Image Watermarking Based on Block SVD and Zernike Moments
Hai Li · 2005
The robustness to geometrical distortions is not well solved in the image watermarking field at present. This paper proposes a robust watermarking algorithm based on block Singular Value Decomposition (SVD) and Zernike moments followed by introducing Zernike moments. The standard image is obtained by preprocessing the original image, then block SVD is applied. The permuted watermark is adaptively embedded into the maximal singular values of each block bit by bit. Watermark extraction is the inverse process of the embedding procedure. Zernike moments are utilized to estimate the rotated angle and then un--rotated to enhance the watermark extraction performance. Experimental results show that the fidelity of the watermarked image is fine and the presented scheme is robust to geometrical distortions such as scaling, rotation and cropping.