Content-adaptive Image Steganography Based On Game-theoretic

Information hiding is one of the important research subjects in the field of current information security.Steganography is a main branch of information hiding and primarily applied to covert communication.It aims to transmit secret information unnoticeably through public channel by hiding secret information into the open digital medium.The converse analytic technology against steganography is called steganalysis,which detects the existence of hidden information,estimates its quantities and extracts it by analyzing the statistical features of cover and stego data.As the two opposite aspects of convert communication,steganography and steganalysis have been developing rapidly in antagonism.Image adaptive steganography is widely regarded as a better method to resist steganalysis.However,the steganalysis algorithms are not specifically designed for adaptive steganography.Therefore,this knowledge does not accord to Kerchhoffs criteria:the attacker has image steganography adaptive rule information or even he may estimate the adaptive rule information.That is to say,steganalyst can base on steganographer’s choice to detect image texture area,thus the adaptive security steganography method would not be high enough.How to design secure steganography has become an important issue under the Kerchhoffs criteria.This paper based on the game theory studies on cover image adaptive steganography.We set up game model by LSB Matching embedding,and analyze both players’ mixed strategy in Nash equilibrium.Then we present the practical application game model in steganographic method for digital images with four-pixel value differencing,and propose a new adaptive steganography algorithm.The achievements are as follows:1)A secondary embedded game model is proposed.We analyze the game theory research on digital image steganography,and point out the weakness of previous game model.Steganalyst will mistakenly identify that no bit is embedded,when the secret bit is the same as the least significant bit of cover image by LSB Matching.Then we introduce secondary embedding and co-occurrence matrix to distinguish cover and stego accurately for improving the steganalysis method of Schottle.We study the relationship of two players in game theory to establish realistic adaptive image steganography game model,and utilize the new Nash equilibrium to guide steganographer choosing security embedded location finally.2)A random non-fully Embedding algorithm is presented.We analyze ordinal and full-embedding method of digital images with four-pixel value differencing,and point out the concealment of this embedding algorithm that is not high enough.That is to say,the distortion of image may cause steganalyst to extract the secret information sequentially.Under the guidance of game theory,we use arnold transform to scrambling the sequence of secret information for the purpose that the steganalyst cannot detect out information orderly and improve the security of steganography.The image system parameters can be obtained,so that the image full capacity is exactly greater than the length of information for non-full embedding.Moreover,these parameters make steganography more flexibility by reducing the distortion of image.