This paper proposes a new blind, robust image watermarking for copyright protection and ownership verification in the DCT domain. The owner information is in the form of a logo image. Two different watermarking systems are considered. The first one is based on the principle of spread spectrum communication, in which we propose an adaptive direct sequence spread spectrum DSSS embedding, and a correlation based decoding. The second one is based on Spread Transform (ST) which is a hybrid combination between spread spectrum and a quantization watermarking scheme called Scalar Costa Scheme (SCS). We describe how to incorporate a perceptual model, based on Watson's perceptual model, into the frameworks. The proposed systems are tested using different standard test images and the performance is evaluated under JPEG compression and compared to each other and to other watermarking schemes. Experimental Results show that the proposed systems have good imperceptibility and higher robustness to JPEG image compression. |
