Parameter optimization of an embedded watermark using a genetic algorithm

A dual image watermarking scheme is improved upon. The dual watermark scheme has the advantage of containing a robust watermark and a fragile watermark. The primary watermark is a PN sequence robust watermark inserted into a fragile watermark which is a secondary image. The latter watermark image is embedded into the host image using watermark embedding factors alpha_i´s. In this study, the values of alpha_i´s are obtained using a genetic algorithm that maximizes the quality of the watermarked image while keeping a high degree of robustness to attacks. When applied on several gray scale images, the watermarked images survived high levels of low pass filtering, compression ratio (90:1) and luminance attacks. The achieved PSNR in case of the Lena image reached a high value of 76 dB.