Digital Image Watermarking Algorithm Based on Texture Masking Model

The trade-off between invisibility and robustness in image watermarking algorithms is considered as one of the major issues in designing watermark-based copyright protection systems. Accordingly, different models had been proposed in the literature to obtain robust watermarked images while maintaining the perceptual quality. However, most of these studies are involved with complex algorithms as using multiple signal transformation tools within hybrid systems. In this paper, a low complexity texture-masking model based on Lifting Wavelet Transform (LWT) is utilized to find the blocks with the highest texture and choose them for watermark embedding. Choosing highly textured places helps to insert the watermark with a further intensity that leads to higher robustness and at the same time the Human Visual System (HVS) is less sensitive to changes in these areas. As a result, high quality watermarked images were produced in terms of objective and subjective evaluations, as the structural similarity value (SSIM) for tested images was larger than 0.99