A New DWT-SVD Based Image Watermarking Technique By Utilizing The Features Of Human Visual System

Image watermarking methodology requires being robust against attacks as well as transparent to human eye. To achieve this goal, this paper presents a new DWT-SVD based image watermarking, where both the cover and watermark images experience DWT-SVD Transformation. In our work, the horizontal subband of first level DWT of host image is used for higher level decomposition. Here, the singular values of second level horizontal coefficients are modified instead of modifying the DWT coefficients directly by the watermark image. Simulation results illustrate that our proposed watermarking scheme can successfully withstand against a variety of image processing attacks with good perceptual quality. Furthermore, it offers the best resistivity and a PSNR more than 40 dB for Lena image