Analysis and detection of text steganographic tool-Stegparty

Using the idea of probability statistics, a study on Stegparty encode-text was done. After analyzing the characteristics of encode-text, a detecting algorithm based on characteristic lexicons was proposed. The idea possesses favorable commonality and flexibility. It also fits other text steganography based on substitution such as synonyms substitution and punctuation substitution. A detection experiment on Stegparty was carried out. Experimental result shows that the detecting algorithm can achieve an accuracy of 90% even when the embedding ratio is very low (about 0.08%).