Title: Critical review of image steganalysis techniques

Authors: J. Anita Christaline; R. Ramesh; D. Vaishali

Addresses: Department of ECE, SRM University, Vadapalani, Chennai, India ' Saveetha Engineering College, Thandalam, Chennai, India ' Department of ECE, SRM University, Vadapalani, Chennai, India

Abstract: Steganography is the art of covert communication aims at hiding information onto a digital cover media. The choice of the cover medium could be audio, image or video files. Owing to the illegal use of such secret communications, steganalysis (the art of code breaking steganography) has gained momentum. This paper presents the recent trends in image steganalysis techniques. The two prominent methods have been identified as embedding specific and universal (blind) steganalysis. As universal (blind) steganalysis does not demand the prior knowledge of the embedding method, it has become the choice of many steganalysers. It has been identified that universal steganalysis is a two class classification problem and could be implemented with statistical analyses or computational intelligence. As this classification depends on the chosen feature set or the image model, this paper has detailed the feature-based image steganalysis techniques. The curse of dimensionality (increased computational time) of the chosen feature sets compel the use of statistical feature reduction methods. Hence, the features set optimisation technique have been reviewed in detail.

Keywords: image steganalysis; classifier; image processing; feature set optimisation; steganography; blind steganalysis; universal steganalysis.

DOI: 10.1504/IJAIP.2015.073715

International Journal of Advanced Intelligence Paradigms, 2015 Vol.7 No.3/4, pp.368 - 381

Received: 08 Nov 2014
Accepted: 01 May 2015

Published online: 16 Dec 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article