Improving Hidden Message Extraction Using LSB Steganalysis Techniques

Abstract

Increased awareness of the role of digital forensics in investigations has led to greater efforts being employed by users to conceal their data, possibly even using algorithms purposely designed to evade detection during steganalysis. A digital investigator seeking to ascertain whether some medium is indeed making use of steganography to hide pertinent evidence must therefore consider including other steganalysis techniques in their analysis in order to overcome the different steganographic strategies that may be used to evade detection. This paper investigates the design of a more comprehensive steganalysis tool that makes use of a series of statistical methods in conjunction with visual and forensic methods to detect messages hidden in images, specifically those hidden in PNG files using Least Significant Bit steganography. The study devises an appropriate combination of the techniques to generate a more effective and comprehensive steganalysis strategy for digital investigators attempting to detect hidden data.