Classification For Computer Graphic Images And Natural Photographs

Along with the development of computer network technology, the information besides the multimedia is transmitted on the internet, more and more secret communication mixed with common data can not be detected. Steganography is the art of covert communication which hides the very presence of communication by embedding covert messages into the other unconfidential cover objects. It has been proved that the terrorist has ever utilized this technique to transfer the secret information and organized the terroristic raid, Consequently, as the steganography transmits the martial, commercial secrets and protects the personal privacy, it has threatened the security of country contemporarily. Currently, the music is prevailed and distributed abundantly on the internet, and it has become one of the primary carriers. This fact also makes the issue in the field of digital audio media's secret corresponding get much attention.Steganalysis is the attack to the steganography. The audio steganalysis aims to reveal the presence of secret messages of audio canier. The work of this paper is exhibited just for this theme. Basing on summing-up of the audio steganography in existence, the detection for LSB is investigated at first. Then the four kinds of universal audio steganalysis methods at present are discussed. Finally, the two detection algorithms for audio-to-image audio steganography are proposed at first time.The details as follows:As prosecuting the detection for LSB, the image's LSB steganalystic ways are extended into audio in this paper, such as Chi-Square test, RS test, SPA analysis. By the experiment, these methods have been reviewed and contrasted in the aspects of detection effect and applicability.For universal audio steganalysis techniques including steganalysis algorithm basing on audio quality metrics, additive noise model, high-order statistic model of audio signal, as well as wavelet transform and linear predictive analysis, the basic principle of them is examined , the advantage and disadvantage of performance for each measure are sumrized, and the expansibility of partial methods are discussed.A new audio steganography method realized by image steganography algorithm—Audio-to-image Transform Audio Steganography is detected in this paper. Through researching the statistic characteristic change of audio signal induced by this measure, two special attacking methods are designed. The first one makes use of the Wavelet Transform, Short-time Fourier Transform, Principal Components Analysis techniques to perform the detective analysis. The second one is by extracting the peak frequency feature which has the linear relationship with the embedding capacity according to the linear regressive model to estimate the length of secret messages. The experimental results prove the efficiency of the proposed algorithms.In the end, we sum up the work of article and discuss the prospect of information hiding.