Sign correlation detector for blind image watermarking in the DCT domain

Digital watermarking is a key technique for protecting intellectual property of digital media. Due to the ability to detect watermark without the original image, blind watermarking is very useful if there are too many images to be authenticated. In this paper, we pose the difference on mathematical models between private watermark detection and blind watermark detection, and then point out the limitation of linear correlation detector (LCD). After reviewing some statistical models which have been proposed to better characterize the DCT coefficients of images, we deduce a new blind watermark detector - sign correlation detector (SCD) based on the Laplacian distribution model. Computing result of asymptotic relative efficiency demonstrates the effectiveness of the detector. A series of experiments show its robustness.