Development of an Intelligent Perceptual Shaping Function in Digital Image Watermarking

In this paper an intelligent method to develop a perceptual shaping function based on DCT_GP has been proposed. To achieve a trade off between two opposed characteristics of digital image watermarking, robustness and imperceptibility, a genetic programming based algorithm has been applied. In this way, the cover work is divided into some blocks and the DCT coefficients in each block are calculated. One AC coefficient, which is not in the low frequency or high frequency parts, in each block is shaped according to a perceptual shaping function. This function is produced by the genetic programming process and can easily be adaptive by watermark and cover image. To evaluate the proposed algorithm, some digital image processing attacks such as image cropping, noise contamination and JPEG compression has been used. Robustness and imperceptibility of proposed algorithm under experimental results, has have improvements in comparison with other similar methods.