Soft computing approach to steganalysis of LSB embedding in digital images

Abstract

Steganography methods embed hidden data in digital images to provide a means of secret communication. Steganalysis methods detect such embedded data based on statistical analysis of a digital image. We experimented with using decision trees for detecting messages hidden in the least significant bit plane of an image and compared their results to those obtained using multilayered feedforward neural networks.