Capacity Investigation of Markov Chain-Based Statistical Text Steganography: Arabic Language Case

Nowadays, there is a high motivation to develop techniques of sending secret information as a result of the need to transfer some data in a secret way. Steganography is one of the important techniques in this field. In particular, statistical text steganography is classified as an immune technique against the majority of steganalysis techniques compared to other steganography techniques. On the other hand, although Arabic is the official language of many countries, many steganography techniques remain unexplored for it. In this work, we implemented a Markov Chain (MC) encoder/decoder combined with Huffman Coding (HC) for Arabic text steganography. Additionally, we computed an upper bound and a lower bound for the stego-text length that depends on the designed encoder/decoder parameters. The proposed Arabic steganography technique capacity performance was investigated for different encoder parameters and secret message length. We found that applying some constraints on the MC increases the embedding capacity till it reaches an upper limit floor.