Adaptive selection of coefficient's portions in the transform domain watermarking

Abstract

This paper investigates the transform domain-based adaptive watermark algorithm based on fuzzy inference. In addition, our approach has double-aim in first to enhance the robustness of the watermarking and in second to reduce the complexity of the fuzzy inference rule. Not only statistical characteristics (SC) but also human visual system (HVS) parameters, such as contrast sensitivity, texture degree and average value are involved to select the optimal coefficients region. Rule based fuzzy inference map is designed by utilizing a possible relation between HVS and SC. The performance analysis of the proposed approach is evaluated by the experiments of imperceptibility and correctness of watermark. Experimental results showed that proposed scheme has a better PSNR and has stronger response than previous results of Cox et al. © Springer-Verlag Berlin Heidelberg 2005.