Compression and watermarking for color images

In this paper a new technique is introduced to hide nine images in a color image. The color image has size four as the watermark size. In each of the color image components three images are hidden. Each color component is decomposed by a Discrete Wavelet Transform (DWT) into four components. Three of four components are used to hide the three images. The hiding operation starts with taking the Discrete Cosine Transform (DCT) for the watermark and the result is multiplied by a factor and added to a component of the three components which result from the DWT decomposition. The Inverse DWT (IDWT) is applied to reconstruct the red, green and blue components. Finally, the result is the reconstruction from the main components. The extraction process is considered in the theory of statistical signal processing as a correlation receiver. The correlation is calculated to decide each pixel is a watermark pixel or not. Different types of images are used as watermarks. The results show the robustness of the technique, the SNR obtained is at least 25 dB.