Mathematical model of steganalysis algorithms correlation estimation with computational complexity considering
Атласов Игорь Викторович, Valerii Vladimirovich Men'shikh, Roman A. Solodukha · 2017
The article is devoted to the construction of a mathematical model and the algorithm for reducing a set of staganalytical correlated methods taking into account the computational complexity of the problem. The basic steganographic methods for the images spatial domain are described. Known steganalytical algorithms for the spatial domain are reviewed. It is shown that a significant number of steganalytical algorithms do not lead to results improvement but slows down the analysis. The necessity of choosing an optimal set of steganalytical algorithms is explaned. The too much power of the set of steganalytical algorithm combinations is shown. The disadvantages of the variance analysis in solving this task are pointed. The theoretical basis for constructing a mathematical model using Fisher and Laplace theorems is presented. The mathematical model of reduction of a set of correlated steganalytical algorithms is given. The peculiarity of the model is the estimation of the computational complexity of the task. The algorithm allowing adding and removing algorithms from consideration without additional calculations is developed.