RGB color image watermarking using discrete cosine transform

Abstract

Recently, access to multimedia data has become much easier due to rapid growth of the internet. Everyone could access these data, and use them for personal or commercial purposes, for these reason copyright problems appeared. Digital watermarking techniques are used to protect the copyrights of multimedia data by embedding secret information in the host media, for example, embedding in images, audios or videos. Many watermarking techniques have been proposed in the literature to solve the copyright violation problems, but most of these techniques failed to satisfy both imperceptibility and robustness requirements. In this thesis, color image watermarking technique is proposed. The proposed technique involves three main stages, which are, pre-processing, embedding, and extraction. In the pre-processing stage, a host image is converted from RGB to YCbCr color space to preserve imperceptibility and robustness, then, Cb component is extracted and partitioned into four quadrants. Finally, canny edges detection algorithm is applied on all quadrants to choose the best quadrant which contains the highest number of edges. Subsequently, in the embedding stage, DCT is applied on the selected quadrant to produce four DCT sub-bands namely LL, HL, LH, and HH. A watermark is embedded in the LL sub-band to obtain a maximum level of robustness. In the extracting stage, DCT is used again to decompose the selected quadrant of watermarked image, and finally the watermark image is extracted. To prove the efficiency of proposed technique, five types of attacks is applied on watermarked image namely, Gaussian noise, Salt & Pepper Noise, Poisson, Speckle and Cropping. The experiments results have shown that the proposed technique successfully withstood against all the mentioned attacks, and at the same time preserved the watermarked image quality