Enhancement of Markov Chain-Based Linguistic Steganography with Binary Encoding for Securing Legal Documents

The Markov Chain-based linguistic steganography algorithm can effectively hide information within human-like cover text, but it is highly limited in processing speed. A traditional implementation relying on Huffman tree-based encoding mainly suffers from slow processing due to the computational overhead of building the tree itself. To address this issue, this study proposes an enhanced algorithm using binary indexing for constant time complexity. The results were experimentally calculated using models of varying state sizes derived from the same text corpus as a control variable. Perplexity analysis was also employed to evaluate imperceptibility and ensure there were no drawbacks to the cover media’s integrity. The results indicate that the enhanced algorithm improves processing speed by up to 54 times across all state sizes without compromising imperceptibility. This establishes that the enhancements yielded a significantly faster processing speed for the existing algorithm while remaining secure in its concealment.  In practice, the algorithm was applied in legal document storage to strengthen its security.