Steganographic Coding And Its Applications In Image Steganography

With the rapid popularity of the Internet and quick development of communication technology, as a new cryptography technique, information hiding becomes a research hotspot in the information security area. As an important branch of information hiding, steganography aims at sending secret message covertly by hiding it into public multimedia data. And steganographic coding is a core technique of steganography, which studies how to embed as many messages as possible per change of the cover object.Targeting on the development and application requirement of steganography, some researches are done on steganographic coding and its applications in the design of image steganographic algorithm in the present thesis. The bodies of the studies are outlined as follows:1. A novel steganographic algorithm based on steganographic software Steghide is proposed. First of all, the ability to preserve the first-order histogram of cover image is improved by using the dynamic minimum degree graphic matching algorithm in graph theory and the maximum modification of cover pixels is reduced to 2. Then the change number of cover pixels is decreased by using binary steganographic codes, i.e. the distortion of cover image is reduced. Furthermore, the maximum modification is reduced to 1 by using the efficient double-layered embedding coding and the embedding efficiency is improved. Finally, the theoretic analysis of the algorithms'embedding efficiency is presented, and the theoretic results are in accordance with the experimental results.2. An edge adaptive steganographic algorithm based on image texture is presented. First, according to the measurement of image texture and the length of secret message, the pixels in the marginal region of cover image are chosen for the embedding of secret message. Then secret message is embedded in these pixels by using the steganographic algorithm based on LSB matching function. The quality of cover image and the ability to resist steganalysis is improved.3. Matrix embedding under general distortion model and its applications are proposed.(1) Under general distortion model, the number of changing patterns rises by increasing the changing number of bits in each block. Then the pattern of the least distortion is chosen, and general matrix embedding is obtained. Modified matrix embedding algorithm is a special example of general matrix embedding.(2) A new measurement of image texture is presented and a new distortion model is defined based on this measurement. Then general matrix embedding is applied to this model. The idea of the steganographic algorithm based on LSB matching function is used to improve general matrix embedding and improved general matrix embedding is obtained. Compared with matrix embedding and general matrix embedding, improved general matrix embedding can have better visual impact of stego image; the total and average distortion is smaller.