Steganalysis of a chaos-based steganographic method

Abstract

Hiding information in a cover image can be regarded as introducing some stego-noise. These alterations may produce some form of degradation or some unusual characteristics and can then be exploited for the sake of detecting the hidden message. In this paper, we analyze the security of a chaos-based steganographic method, namely the Chaotic Edge adaptive Least Significant Bit Matching Revisited (Chaotic EA-LSBMR). We focus on the universal image steganalysis method which uses the higher-order statistics as features taken from the wavelet multiresolution approach. A Fisher Linear Discriminant classifier is used to classify cover and stego images.