The Research On Natural Language Information Hiding Technology Based On Steganography Coding

Information hiding technique makes use of the redundancy of human sensing system and carrier data to embed secret information into image, audio, video, text and other digital media for covert communication. In pepole's daliy life, the most frequently used tool of information communication is language and words. Therefore, information hiding technology using natural language text has attracted many people, but most algorithms were of poor ability of anti statistical analysis. Aiming at sloving this problem, this paper studied natural language steganography technology based on word layer and proposed novel steganography coding algorithms to improve security proformance of existing synonym substitution-based steganography algorithm. The main research contents of this paper include the following aspects:1. Embedding secret information using general synonym substitution-based method would result in low security performance and high rewrite rate as it makes a lot of modifications on a carrier text. A new hiding algorithm combining matrix coding with synonym substitution was proposed to deal with this drawback. In order to embed secret information, matrix coding was applied according to secret information for grouping synonyms in the cover text to determine the position of synonym needed to modify. Lower rewrite rate and higher security peoformance were achieved by uing matrix coding. Experimental results show that this algorithm improved embedding efficiency and reduced the possibility of stego texts to be detected by statistical analysis, also guarantee the security of secret information.2. Traditional synonym substitution information hiding method changed the frequency of synonymy in the text and destroyed the features of the original text, providing the chance for the detection system using word frequency analysis to find the existence of secret information. In the view of this shortcoming, this paper proposed a steganography method combined multiple-base coding with word frequency. By maintaining the number of low frequency words in each group during information embedding, the number of low-frequency synonyms were invariant before and after substitution. The experimental results show that the algorithm can resist the attack from the detection algorithm based on synonym word frequency analysis.