A steganographic method based on self-organizing map and HVS

In order to provide large capacity of hidden secret data and imperceptibility of stego-image, in this paper, a novel image steganographic method based on self-organizing map and HVS was presented. According to contrast and texture sensitivity, self-organizing map based on competitive learning is trained. So NNs trained is the key of the embedded and extracted secret data. The method exploits neighboring pixels(upper, left, right, bottom) to estimate the degree of sensitivity of pixels with NNs trained so that pixels in less sensitive areas can potentially carry more hidden data. From the experimental results, compared with SOC, the proposed method can hide a much larger information and maintains a better visual quality of stego-image.