Steganalysis Using a Composite Set of Transform Domain Features and Ensemble Classifier

Abstract

In this paper, generic as well as analytic steganalysis method is used to detect the presence of hidden image and also the tool used to embed the secret image in the cover image. Identifying the type of embedding algorithm might lead to extraction of the hidden image. A new set of features including Noise features, Zernike moments, Moments of Characteristic Function (MOCF), Colour, Fourier Descriptors are extracted independently and are given to different classifiers including Minimum Distance Classifier (MDC), Least Squares Support Vector Machine (LS SVM), OSU SVM etc and their steganalytic performance is analysed. In order to improve their performance, ensemble classifier is used. The performance of the Steganalyzer is also analyzed with different number of training, testing subjects. The experimental results demonstrated that the proposed method can effectively identify digital images from their tampered or stego versions and can also successfully classify the steganographic algorithm with which the secret was embedded into the cover image.