A geometrically robust watermarking scheme based on perceptual hashes and genetic Algorithm

A geometrically robust image watermarking scheme based on perceptual hashes and genetic algorithm is proposed. The synchronization information, which is generated by perceptual image hashes and pseudo-random number (PRN), and the watermark are embedded into the image based on the human visual system (HVS). In watermark extracting, the geometric transformation parameters are computed from perceptual hash values of the image, which is realized by genetic algorithm. And then the image is re-synchronized by inverse transforming with the parameters. According to the locally optimum detection and a threshold can decide the watermark is present or not, and the threshold is obtained by the Neyman-Pearson (NP) criterion. Simulation results show that the proposed scheme is not only invisible, but also robust against regular signal processing and geometric attacks.