Hiding Information Detection Technology Research Based On Still Digital Image

Illegal implementations of steganography grievously damage the security of internet. To supervise the use of steganography and prevent the illegal uses has been a main part of national security. Steganalysis has been a major research content to solve the security problem effectively. Detection technology is the basis of steganalysis to inspect objective documents with secret information or not, it is the most important part in steganalysis. In this paper hidden information-detect technology based on images was discussed; the main work was summarized as follows:1) Analyze a general steganography model, combining with characteristic of steganalysis, this paper improve the attack model, divide the attacker into positive attacker, positive steganalysis and passive steganalysis.We use formalization method to describe the connection between steganography and steganalysis,then expound the key point of steganalysis.2) Analyze the characteristic of some typical steganalysis algorithms. Discuss the effect of algorithm statistic features, the influences of image noise, and the reason for low efficiency of steganalysis algorithms. Analyze the influences to steganalysis by image noises based on grads energy. Then propose a method to reduce the influence.3) Propose a new multi-feature steganalysis method based on the correlation of image bit planes. Collect the statistic quantities on the base of above correlation and common traits of replace algorithms in spatial domain. And detect this type of algorithm by multi-feature steganalysis technology effectively.4) Propose a new data hiding domain recognize-method based on statistic features. Based on the analysis of carry-image changes in different hiding domain, we use multi-classifier to recognize the data hiding domain by mulite-feature technology. This section geive a technology support to hidden imformation extract technology, which make steganalysis effectively.Finally, we summarize our research work for the thesis, and discuss further research topics and directions in the future.