Advertisement
Published in:
2011 | OriginalPaper | Chapter
LinL:Lost in n-best List
Authors : Peng Meng, Yun-Qing Shi, Liusheng Huang, Zhili Chen, Wei Yang, Abdelrahman Desoky
Published in: Information Hiding
Publisher: Springer Berlin Heidelberg
Activate our intelligent search to find suitable subject content or patents.
Select sections of text to find matching patents with Artificial Intelligence. powered by
Select sections of text to find additional relevant content using AI-assisted search. powered by
Translation-based steganography (TBS) is a new kind of text steganographic scheme. However, contemporary TBS methods are vulnerable to statistical attacks. Differently, this paper presents a novel TBS, namely Lost in n-best List, abbreviated as LinL, that is resilient against the current statistical attacks. LinL employs only one Statistical Machine Translator (SMT) in the encoding process which selects one of the n-best list of each cover text sentence in order to camouflage messages in stegotext. The presented theoretical analysis demonstrates that there is a classification accuracy upper bound between normal translated text and the stegotext. When the text size is 1000 sentences, the theoretical maximum classification accuracy is about 60%. The experiment results also show current steganalysis methods cannot detect LinL.