A hybrid watermarking scheme with CS theory for security of multimedia data

Abstract

A hybrid watermarking scheme for multimedia data such as digital image, digital video is proposed in this paper. The scheme utilizes various image processing transforms and Compressive Sensing (CS) to achieved fragility and security for multimedia data. The compressive sensing is applied on the singular value of wavelet coefficients of watermark image to get the CS measurements. These CS measurements are embedded with embedding factor into the hybrid coefficients (high frequency curvelet coefficients of DT coefficients) of the host medium. In the proposed scheme, host medium may be digital image or video. The generated watermarked data have different security layers such as four levels from CS such as knowledge of transform basis matrix, measurement matrix, orthogonal matrices, information of wavelet basis matrix and one level from watermarking such as embedding factor. The scheme offers fragility and security to multimedia data against various standard watermarking attacks such as signal processing attacks, geometric attacks, and compression attacks. This scheme can be provided high payload capacity and used for host multimedia authentication. A comparison of proposed scheme with existed schemes shows that the proposed scheme performed faster and better than existed schemes available in the literature.