Incorporating HVS parameter into the transform-based watermarking
Yoon-Ho Kim, Heau-Jo Kang · 2006
In this paper, transform-based watermarking scheme based on human visual system is proposed. Human visual parameters are useful to ensure the imperceptibility of watermark image. Not only human visual parameters, such as contrast sensitivity, texture degree but also statistical characteristics are involved to select the optimal coefficients region. The performance of proposed algorithm is evaluated by the experiments of imperceptibility and correctness of watermark. According to some experimental results, contrast sensitivity function is superior in smooth image. On the other hand, statistical characteristics provide good results in rough images. Consequently, how to select the parameters considering image attribute is key problem in effective watermarking.