IMAGE STEGANOGRAPHY IN COLOR CONVERSION

For image steganography, no matter how ingenious the distortion function is designed, the embedding operation will lead to changes in the distribution of cover image, which causes security risks. In this brief,an image steganography based on de-colorization and colorization is proposed, which can accomplish the embedding and extraction task of secret information during the process of color translation.Unliketraditional image steganography, the proposed steganography counteracts the embedding influence of secret information through de- colorization network. In addition, a robust embedding algorithm is utilized to embed secret image into cover image by considering the de-colorization and colorization as an attack. Moreover, different training datasets are created to analysis their impact on the reconstruction performance of the colorization network. Finally, simulation results are given to verify the feasibility of the proposed steganography.