Data Hiding Based on Intelligent Optimized Edges for Secure Multimedia Communication

Recently, image steganography has received a lot of attention as it enables for secure multimedia communication. Payload capacity and stego image imperceptibility are a critical factors of any steganographic technique. In order to receive maximum embedding capacity with a minimum degradation of stego images, secret data should be embedded carefully in a specific regions. In this paper, data hiding is considered as an optimization problem related to achieving optimum embedding level of the cover image. Embedding data in edge area provide high imperceptibility. However, the embedding capacity of edge region is very limited. The work attempt to improve the edge based steganography by incorporates edge detection and vision science research. Genetic Algorithm that uses human visual system characteristics approach for data hiding is presented. Primarily, the approach applies Differences of Gaussian detector which closely resembles the human visual behavior. Secondly, the edge profusion indicates the level of threshold visibility with the help of Genetic Algorithm training. The suggested solution uses Contrast Sensitivity Function (CSF) which produces the edges based on the size of the embedding information. The authors of this paper compared their technique with other classical and recent works. The quality of the steganography is measured based on various quality metrics such as PSNR, wPSNR, SSIM and UIQI. These metrics declare the stability between imperceptibility and large embedding capacit