Efficient Genetic Algorithm Based Image Watermarking Using DWT-SVD Techniques

This paper presents a new method to improve the robustness and imperceptibleness in watermarking. In this, the singular value of watermark is embedded to singular value of 3rd-level DWT approximation matrix of original image. The genetic algorithm is used to optimise the scaling factor with which the watermark is embedded to host image. The algorithm makes use of fitness function that takes two values PSNR and correlation. Here we also find the transparency and robustness of watermark under various attacks.