A Steganography on Synonym Frequency Distribution

Abstract

Linguistic steganography is concerned with hiding secret information in natural language text imperceptibly. One of the major methods for doing so is synonym substitution. Existing schemes hardly consider the influence on word frequency so that they can’t resist the text steganalysis which bases on synonym frequency. This study takes the features of word frequency into account, and proposes a novel method using a combination of synonym substitution and multiple-base. This method mainly consists of three steps. Firstly, synonym database is to be preprocessed according to synonym frequency. Then, the secret data will be transformed into the expression of multiple-base expressions. Finally, the chosen synonym will be used to hide the secret message according to its continuous dimension and the proposed algorithm. In this paper, continuous dimension is a new concept to be proposed, and it is important to the embedding and extracting processes. The proposed method possesses superiority to other related methods in terms of security and robustness. Moreover, the most important goal in this method is trying to remain the synonym frequency distribution and achieve good imperceptibility. This paper describes in detail the steganography algorithm that involves the embedding and extracting processes, and presents its advantages and limitations.