When designing a watermarking algorithm, trade-offs exist among three parameters: payload, fidelity, and robustness. Data payload is the number of bits that can be embedded in the digital data; the fidelity is the degradation introduced into the signal; and the robustness is the ability of the watermark to remain readable after innocent or malicious signal processing operations on the watermarked image. These parameters are conflicting with each other, and they should be set to meet the requirements of the application (Levicky & Foris,2004).
Watermarking techniques are generally classified into spatial domain or the transform domain. The earlier watermarking techniques were almost in the spatial domain. Spatial domain techniques are not resilient enough to image compression and other image processing (Potdar & et al, 2005). Although some methods, e.g. in (Depovere & et al,1998), adopted pre-filtering skill to increase the percentage of identification, experimental results have shown the fundamental disadvantages of spatial domain watermarking. Transform domain watermarking schemes like those based on the discrete cosine transform (DCT) ((Chu, 2003), ( Lin & Chin2000), ( Deng & Wang, 2003)) and the discrete wavelet transform (DWT) ((Hsieh &et al, 2001), ( Reddy & Chatterji, 2005), (Tay & Havlicek, 2002)) typically provide higher image imperceptibility and are much more robust to image manipulations.
It is generally believed that the performance of most existing watermarking systems is not close enough to the fundamental limit on robust watermark embedding rates at which high perceptual image quality is maintained. Although, embedding the watermark in the perceptually significant coefficients could alter the perceived visual quality of the image. Thus, two essential prerequisites for a powerful watermarking scheme, robustness and invisibility do conflict with each other (Amirgholipour & Naghsh-Nilchi, 2009). To support real applications demanding imperceptible and robust watermarking, image preprocessing techniques are required. Any watermarking scheme consists of two parts: Insertion algorithm and extraction algorithm. Robustness of watermarking algorithms could increase by improving each of these parts. Based on this fact, different techniques are applied to improve imperceptibility and robustness of existing watermarking algorithms (Amirgholipour & et al, 2009).
Most of these techniques are based on preprocessing multimedia data before embedding watermark information. They are concentrated on finding the most proper places in the multimedia to embed watermark. Methods such as employing JND profile, contrast sensitivity and texture masking, and exploiting characteristics of the human visual system (HVS) are used to embed watermark in places which are perceptually invisible and more robust (Lahouari & Bouridane, 2005). The goal of these methods is to develop an efficient, yet simple, perceptual model based on a sub-band decomposition that is...