Feasibility of text visualization in text steganalysis

• Published in: Frontiers in Artificial Intelligence and Applications
• Main Author: Ab. Rauf R.H.; Jamal N.
• Format: Conference paper
• Language: English
• Published: IOS Press BV 2014
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-84948776300&doi=10.3233%2f978-1-61499-434-3-103&partnerID=40&md5=895ad9c23a80cefcfd9265d45cb8abde
• ISSN: 9226389
• DOI: 10.3233/978-1-61499-434-3-103

Text steganography was never a threat until it has been manipulated by cyber criminals or terrorist for their own benefit. Secret messages sent through steganography with bad intent can cause harm and to the extent jeopardizing the security of a company or country. Thus, text steganalysis is critically needed for forensic investigation purposes due to the tremendous usage of digital media in accessing and disseminating information. Apparently, there has not been much research done on text steganalysis and mostly researches are based on statistical approach. The aim of the research is to propose a new text steganalysis technique in detecting format-based steganography using a simple and effective approach. This paper introduces a novel text steganalysis technique based on color-coded text visualization. An encoding scheme for the text visualization is designed by analyzing text features with respect to colors to detect whitespace pattern. The aim here is to distinguish between natural and stegano text by color-coded visualization. Experiments show that the detection performance accuracy successfully reaches 96.67% with remarkably high precision and recall. This finding has proved that the text visualization technique is capable in detecting text steganography effectively by simply looking at the text visualization image. It is evidently shown that a simple text steganalysis technique is successfully discovered and it is feasible. © 2014 The authors and IOS Press. All rights reserved.