Fuzzy Comprehensive Evaluation Of Steganalysis Algorithms

Steganography is the art of secret communication. Its purpose is to hide the very presence of communication. And, steganalysis is the process of detecting steganography by looking at variances between bit patterns and unusually large file sizes. It is the art of discovering and rendering useless covert messages. Now, a lot of steganalysis algorithms have been proposed, but it is difficult to choose a proper one for actual applications. And, the evaluation and selection of steganalysis algorithms is a problem with multi-hierarchy and multi-objection. So fuzzy comprehensive evaluation has great significance to the selection of steganalysis algorithms. Aiming at meeting this practical requirement, this thesis is focused on the research of fuzzy comprehensive evaluation model of steganalysis algorithms on the base of the features of steganalysis algorithms and the theory of fuzzy comprehensive evaluation algorithm. In this model, it uses the reliability and the speed of the steganalysis algorithm as the first level factors. As the reliability, it has the second level factors and the third level factors to indicate the influence of the different embedding methods and the different rate of embedded message. As far as subject functions of basic factors of reliability are concerned, by analyzing the collected test data, this thesis presents that the distributions of detecting results are like the normal distribution. So the subject functions are constructed by using this distributed function. For the subject function of speed, triangular fuzzy number is used to construct the function. And the model also uses the AHP method to get the weight of the factors of the steganalysis algorithm.