A Steganalysis Method Based on the Distribution of First Letters of Words

Semantic steganograhpy is a method using texts to hide secret messages, which embeds messages into texts at the level of linguistic understanding. This paper studies the semantic steganography method, and analyzes the statistical features to search the differences between stego-texts and natural texts by modeling the latter. And a steganalysis method is designed based on the distribution of first letters of words according to the differences. The method is easy to achieve, and experiments prove the efficiency and reliability of the method.