Implementation of Radial Basis Function Neural Network for Image Steganalysis

摘要

Steganographic tools and techniques are becoming more potential andwidespread. Illegal use of steganography poses serious challenges to the lawenforcement agencies. Limited work has been carried out on supervisedsteganalysis using neural network as a classifier. We present a combinedmethod of identifying the presence of covert information in a carrier image usingfisher?￠a??a?￠s linear discriminant (FLD) function followed by the radial basis function(RBF). Experiments show promising results when compared to the existingsupervised steganalysis methods, but arranging the retrieved information is still achallenging problem.