A Coverless Plain Text Steganography Based on Character Features

Abstract

The coverless text steganography has become a research hotspot in the field of information hiding because no modification to a text carrier will improve the concealment of steganography. However, compared with images, audios, videos, or other carriers, a plain text has less redundancy for modification, so the text steganographic embedding rate is usually lower, and semantic anomalies are more easily produced once it is modified or generated. This paper analyzes the parity of Chinese characters’ stroke number and its statistical characteristic, and reach the conclusion that this parity feature exhibits a uniform distribution and it exists in every character. These indicate that the parity can be employed to express a binary digit and its combination can represent a binary digit string. In the light of the above analyses, we propose a new coverless plain text steganography based on the parity of Chinese characters’ stroke number and extend it to further improve the embedding rate. By exploiting the space mapping concept, this method initializes a binary search tree based on the texts from the Internet and searches the corresponding texts according to the secret binary digit string generated from a secret. Thus, a secret will be mapped into a list of the corresponding uniform resource locators. That is, our proposed methods only employ the existing uniform resource locators to convey a secret. Compared with the existing text steganography techniques, our methods not only have high-embedding rate and good concealment but also can improve the embedding capacity without impact on the concealment, which eliminates the contradiction between the embedding capacity/rate and security, because no modification or additional information is generated. In addition, these methods have a prominent search success rate, and they can be extended to other language features as well as to other languages.