Watermarking of Color Images based on a multi-layer process

In this paper we present a new method of watermarking devoted to color images. Watermarking technics generally sign images by introducing changes that are imperceptible to the human eye, but are easily recoverable by a computer program. The locations in the image where the signature is embedded are determined by a secret key in order to prevent possible attacks by pirates or alterations due to compression and coding transformations, or geometrical transformations. Meanwhile a lot of methods have been proposed to watermark grey level images, only few methods have been devoted to color images. In [1], Kutter proposed a new method to watermark a color image based on amplitude modulation of the blue channel, depending on the value of the bit, and proportional to the luminance. Next, Kim et al. proposed in [2] an another method based on magnitude increase of the saturation component with the contrainst that the resulting color difference is acceptable to the human visual system (HV S). Likewise, Fleet and Heeger proposed in [3] to use a human color vision model to control the amplitude of color degradations in order to ensure the invisibility of the embedded signal. More recently, Colduc and al. proposed in [4] an histogram embedding strategy based on a region selection approach to watermark the luminance component. Lastly, Vidal et al. proposed in [5], and Campisi and al. proposed in [6], an another kind of strategy which consists to watermark the image in the frequency domain. The main limit of these methods is that they embed only one data feature. Thus, in [1] Kutter embeds the blue component into the spatial domain, in [2] Kim embeds the saturation component into the spatial domain, and in [3] Fleet embeds the yellow-blue component into the frequency domain. Rather than embedding only one data feature, we propose a new strategy which consists to embed : the luminance (Y ) channel at many different pixel locations in the spatial domain relatively to luminance values, the saturation (S) component of the (CrCb) chromatic plane at many different pixel locations in the spatial domain relatively to the saturation values, the luminance (Y ) channel at many different pixel locations in the spatial domain relatively to chromatic edge values, the color (Y CrCb) components at many different pixel locations in the spatial domain relatively to the Contrast Sensitivity Function (CSF), the luminance histogram (HY ) of each regions in the spatial domain relatively to regional emergence of regions, the saturation (S) component of each regions in the spatial domain relatively to their hue.