Countering Steganographic Security with CNN

Abstract

Accurate steganalysis is required for extracting secret data hidden using different steganographic techniques. A steganalysis technique using the "Yedroudj" network is proposed in this work. The common steganalysis process is followed using highlight extractors and classifiers. The suspected images are classified with Convolutional Neural Network(CNN) and tested. The accuracy of detecting a steganography image with the proposed network and strategy was around 86.36%. Extracting data from images that used steganography tools like Xiao and Open puff are tested using the yedroudj network by detecting the steganography information.