Watermarking Scheme based on Singular Value Decomposition and Homomorphic Transform

A semi-blind watermarking scheme based on singular-value-decomposition (SVD) and homomorphic transform is proposed. This scheme ensures the digital security of an eight bit gray scale image by inserting an invisible eight bit gray scale watermark into it. The key approach of the scheme is to apply the homomorphic transform on the host image to obtain its reflectance component. The watermark is embedded into the singular values that are obtained by applying the singular value decomposition on the reflectance component. Peak-signal-to-noise-ratio (PSNR), normalized-correlation-coefficient (NCC) and mean-structuralsimilarity- index-measure (MSSIM) are used to evaluate the performance of the scheme. Invisibility of watermark is ensured by visual inspection and high value of PSNR of watermarked images. Presence of watermark is ensured by visual inspection and high values of NCC and MSSIM of extracted watermarks. Robustness of the scheme is verified by high values of NCC and MSSIM for attacked watermarked images.