A new perceptual image watermarking technique based on adaptive fuzzy clustering

In this paper, a new perceptual image watermarking technique in an adaptive fuzzy clustering framework is developed. In this technique, the host image is first decomposed into DCT blocks, which are then classified efficiently using an adaptive fuzzy classification scheme. The reference watermark is then perceptually embedded into each class of blocks separately, thus increasing robustness to attacks. For watermark detection, the watermarks embedded in different classes are extracted and combined together to estimate a reliable watermark. Experimental simulations are provided to demonstrate the efficient performance of the developed technique in terms of robustness to a variety of attacks and reliability of watermark detection.