Information Hiding And Detecting On Image Frequence Domain

Information hiding has become a hot interest area in the research of information security in the internet age. Information hiding is the technique used for covert communication. The detection to hiding is the technique to detect the hidden message and destroy the covert communication. It is very important to research the hiding technique for protecting the communication security, while the research on detection technique is very important to prevent confidential message revealed through hidden cover and avoid the covert communication. Based on the wide used image hiding, in terms of the transform of image and the transform coefficient model, the research has been taken deeply. Refering to hiding detection: the detection algorithm to image transform domain hiding has been researched and proposed. Refering to image transform domain hiding: the F5 algorithm in DCT domain has been improved and a secure hiding algorithm in DWT domain has been researched and proposed. The main contributions of this thesis are summarized as follows:1. Research on the detection to image transform domain hidingThe theory of transform domain has been researched. Based on the analysis of coefficient model with respective to image DWT and DCT, the change of the model parameter after hiding is deduced, and the method of embedding again has been proved. Then the model parameter and randomness of coefficients in image transform domains are extracted, to which the method of embedding again is applied to generate the feature. Finally, the non-linear support vector machine is used for classification, and a effective detection to image transform domain hiding is realized.2. F5 algorithm analysis and modificationBy means of the histogram characteristics of DCT AC coefficients, the F5 algorithm and its detection algorithm are analyzed. The shrinkage of histogram of F5 results in the effective detection. Therefore, a method can eliminate the shrinkage has been advocated and proved in theory, and then F5 is improve based on the method. The histogram preservation of improved F5 is analzed and the detection to F5 is used to attack the improved algorithm. The results of experiment show the improved method can can eliminate the shrinkage and get better security.3. The research on image DWT domain hiding algorithmA DWT domain hiding algorithm based on matrix encoding and statistics preservation is analyzed and devised. Matrix encoding is adopted to eliminate the modification to cover image, and in order to resist the detection, parity representation is used to denote the hidden message. Moreover, the high frequency coefficients are revised to preserve the characteristic property, which makes the algorithm good imperceptibility and statistics preservation. What's more, the embedding locations random selection is utilize to enhance the anti-extracting of algorithm. The analysis and experiment to the algorithm show that good statistics preservation has been got and a perfect resistance to transform domain detection has been obtained.Finally, the research topics are summarized; the further research topics about image transform domain hiding and detection techniques are outlined, and the research directions in future are discussed.