Reversible Watermarking Scheme Using PVD-DE

Abstract

A new reversible watermarking scheme has been proposed using PVD and DE scheme. Here, two copies of watermarked images are generated after embedding watermark within cover work. The logo image is used as watermark for image authentication and message digest calculated from cover image is used for tampered detection. Four bits authentication code are embedded using Pixel Value Difference (PVD) within two consecutive pixels and one bit tampered detection code is embedded using Difference Expansion (DE) method within same pixel pair. After embedding watermark, two sets of watermarked pixel pair has been generated. Now, we distribute two watermarked pixel pair among dual watermarked image depending on a shared secret key. At the receiver end the authentication code and message digest are successfully extracted using PVD and DE from dual watermarked image. The scheme is reversible because it can recover original cover image from dual watermarked image after extraction the watermark. Finally, it has been compared with existing state-of-the-art methods and obtained reasonable better results in terms of capacity and quality measured by bpp (bits per pixel) and PSNR (Peak signal to noise ratio) respectively.