Statistical Restoration to Resist Chi-square AttackBased on Heuristic Algorithms in SteganalyticSystems

Steganalysis are used to detect whether an imagecontains a hidden message. By analyzing different image featuresbetween stego images and cover-images, a steganalyzer is able todetect stego images. In this paper, we present a new method inLSB embedding by avoiding the change of statistic features. Afterembedding data in the first LSB bits, we apply Particle SwarmOptimization by adjusting the second LSB bits of a stego imagewhile creating the desired statistic features to generate themodified stego images that can break the inspection of chi-squareattack. Experimental results show that our algorithm can notonly pass the detection of chi-square test, but also leave ourhidden message unchanged on the first LSB bit, and enhance thepeak signal-to-noise ratio of stego images. By using PSOalgorithm, speed of convergence also is improved

کلیدواژه ها:

Steganalysis ، Statistical Restoration ، PSO algorithm