Digital Watermarking based on Non-Sampled Contourlet Transform

In this paper, a novel digital watermarking algorithm based on non-sampled contourlet transform (NSCT) is proposed. This algorithm gives a model of human visual system (HVS) in NSCT domain, exploiting the masking characteristics of HVS to embed watermark adaptively. NSCT has the characteristics of multi-resolution, localization, multi-directionality and anisotropy. Thus NSCT is much more similar to the information processing mechanism of the HVS than wavelet transform. Additionally, NSCT is a redundant transform, which can increase robustness to additive noise processing, and watermark can be considered as noise, it can be assumed that this algorithm can outperform the algorithm in wavelet domain. This assume is confirmed by experiments , compared to the algorithm in wavelet domain, this algorithm can embed much more watermark data, and the watermark is much more robust, achieved the best compromise between robustness and invisibility.