Improved Spread Spectrum Image Watermarking Via Difference Modulation

Currently, spread spectrum watermarking is a kind of robust watermarking technology. How to improve the robustness of spread spectrum watermarking becomes the focus of current research on robust watermarking. In this paper,we use the differences between the pixel value and its predictive value, which is calculated by its surrounding pixels, to embed watermark so that we can reduce the effect by the original image on the robustness, and then strengthen the robustness of spread spectrum watermarking. Theoretical analysis shows that our algorithm can achieve small false alarm and false rejection probabilities simultaneously, which implies that it has good robustness. Experimental results demonstrate that the algorithm is robust to common typical attacks such as JPEG compression, median filtering, noise addition and image resizing.