Steganalysis of LSB Embedded Images Using Variable Threshold Color Pair Analysis

Abstract

Steganography and steganalysis are important areas of research in recent years involving a number of applications. Steganography is the science of embedding information into the cover image without causing statistically significant perturbations to the cover image. Steganalysis is the technology that attempts to defeat steganography by detecting the hidden information and extracting or destroying it. In this paper we present twenty four bits BMP image color pair analysis with variable threshold (CPAVT) to detect stego-object with 10% payload. In earlier works 20% payload was used in close color pair analysis. It is observed that with new variable threshold technique, the performance parameters, i.e. false detection rate (FDR) and false alarm rate (FAR) are better in comparison with the earlier works.