Robust image watermarking with Zernike moments

Abstract

In this paper, we studied the image watermarking system based on Zernike moments. Because the magnitudes of Zernike moments have the desirable invariance property against image rotation, we employ Zernike moments as our watermark signal to achieve the robustness against image rotation. The watermark is embedded to the cover image in the spatial domain. Based on whether the image is RST attacked or not, two watermark detection algorithms are proposed and tested. The first algorithm can reconstruct the embedded watermark with the Zernike moments' vector extracted from the testing image if it does not undergo any RST attacks. The second algorithm extracts the feature vector in the format of the magnitudes of the Zernike moments from the testing image and employs the RMSE to detect the watermark