A Robust Image Watermarking Scheme Based on the Relationship of SVD

In this paper we propose a novel invisible digital watermarking scheme based on singular value decomposition (SVD). In the proposed method, we first divide the image into 8x8 sub blocks, and then each block is decomposed by SVD. The second and third singular values of each block are exchanged if the watermark bit is 1, so the order of the SVs is not in the original descending order any more. This change of relationship of the order will be detected in the extracting process. The experimental results show that the proposed scheme is robust against common image processing operations, such as JPEG compression, additive Gaussian noise, and median filter.