High Capacity Information Hiding Algorithm For Images

With the rapid development of Internet and communications, information hiding technology has become a hot topic in the filed of information security. Steganography is an important branch of information hiding technology. By using the redundancy of digital carrier, secret messages is embedded into images, audio, video and other multimedia files, without changing the characteristics of the carrier. Thus, the human eye can not distinguish the difference between cover images and stego-images. As a visual form of multimedia data, images are widely used in the internet. The large capacity steganography based on digital images is studied in this paper.The main work includes the following respects:1) The Pixel-Value Difference steganography has poor visibility and a low embedding capacity. In order to solve these problems, a steganographic method is improved in this paper. Firstly, the cover image is partitioned into non-overlapping blocks of two consecutive pixels. Then the difference value is calculated from the values of the two pixels in each block. In this paper, all difference values are classified into a number of ranges based on the characteristics of human visual system. The difference value of each pixel block belongs to a range, and the number of secret bits that can be embedded in a pixel block is decided by the range. Thus, the number of embedded bits in smooth areas is more than that in edge areas. By the introduction of the pixel adjustment factor, the modified amplitude of carrier image is reduced in this paper. All carrier pixels can be used to embed secret information by adjusting the out-bounds pixels. Experimental results have demonstrated that the improved scheme has good security and high capacity.2) The embedding capacity of EMD(Exploiting Modification Direction) steganography is low. In order to improve the embedding capacity, a high capacity steganographic method based on EMD is designed in this paper. Firstly, for the purpose of selecting carrier pixels randomly, the cover image is scrambled. After that, the pixels are grouped. In the process of embedding algorithm, the length of each piece of secret message is adaptive. Each piece of secret message is carried by one group of pixel. Finally, only one pixel or two pixels are increased or decreased by1. The improved method provides four different modified directions, and all values within the modified range are fully exploited. The experiment results have demonstrated that the improved method has high embedding capacity and good quality. This algorithm is robust to RS attacks and has a good performance in PSNR(Peak Signal to Noise Ratio) and SSIM(Structural Similarity index).