Research On Image Steganalysis Algorighms

Information hiding techniques started to rise since the beginning of the 1990s. With the emergence and broad spread of various kinds of information hiding algorithms and tools, it is becoming a hot spot in information security technical field how to effectively control the legal applications of the information hiding techniques. Digital image is a very important storage and transmission format of multimedia information on the Internet. Therefore, images become a very suitable carrier for information hiding techniques because of their widely use and large redundancy space. Information hiding techniques based on digital images, also called image steganography, is always positioning in the center of the whole information hiding research system and draws the attention from a lot of related research institutions and organizations. Meanwhile, image steganalysis, as a counterpart of image steganography, is becoming a new focus in the information security technical field.The paper first introduces the concept, model and development of information hiding techniques briefly and then introduces several categories of image steganography algorithms and corresponding steganalysis algorithms in detail. Then, the paper focuses on analyzing the system model of image steganography based on additive noises, thoroughly discusses the use of corresponding steganography algorithms and tools and then proposed a new steganalysis algorithm for detection. Taking advantage of the independence between the cover image and the secret message and the sparseness representation of the signals, the new algorithm constructed a three-dimensional scatter diagram to illustrate and describe the effect of the embedding process to the histogram and adjacent pixel difference histogram of the cover image, and designed an efficient parameter from the diagram to measure the change of the statistical character of the cover image. Then, the proposed steganalysis algorithm is simulated and a series of standard testing images are tested to verify the effectiveness of the algorithm. Finally the performance of the algorithm is analyzed based on the testing results and the way to improve the performance of the algorithm is put forward.