Generation Method of Dynamic Coverless Arabic Text Information Hiding Using First-Order Markov Chain

In recent years, information hiding has become a crucial aspect of secure communication and data protection. The authors of this study offer a new coverless text information concealment approach based on FirstOrder Markov Chain. The proposed method consists of three steps: the first-order Markov Chain construction, the hiding process, and the extracting process. This method employs a Markov Chain model to determine the frequency of each word or token in the training set, which is then used to generate the transition probability and state transition diagram. Sort the words in descending order based on their branches and possibilities. Configure the code word based on the number of branches. In the hiding process, each character in a secret message token was translated to binary format. The stego-text (that sent to the receiver) was generated by successfully matching the code word. Throughout the extraction procedure, a dynamic implementation of the previously agreed-upon construction first-order Markov chain was employed to extract hidden messages. Three Arabic datasets are used in this work (SANAD (Single-Label Arabic News Articles Dataset) includes 45500 articles, the Arabic Poem Comprehensive Dataset (APCD) contains 1,831,770 poetic verses in total, the Arabic Poetry Dataset contains more than 58000 poems). Because it is unmodified, the proposed method resists current detection techniques since it is unaltered. Furthermore, the hiding capacity has been raised to 5.5, and perplexity has decreased to 14.5. Based on the frequency of words in the Arabic dataset, the success rate is close to 100%, which means enabling them to hide any message successfully.