An Image Steganalysis Algorithm Based on Rotation Forest Transformation and Multiple Classifiers Ensemble

Abstract

In order to enhance the detection rate of ensemble classifiers in steganalysis, concern the problems that the accuracy of basic classifier is low and the kind of basic classifier is single in typical ensemble classifiers, an algorithm based on rotating forest transformation and multiple classifiers ensemble is proposed. First, some feature subsets generated randomly merger with training sample to generate sample subsets, then the sample subset is transformed by rotating forest algorithm and train some basic classifiers, which is made of fisher linear discriminate, extreme learning machine and support vector machine with weighted voting. At last, the majority voting method is used to integrate the decisions of base classifiers. Experimental results show that by different steganography approaches and in different embedding rate conditions, the error rate of proposed method decreased by 3.2% and 1.1% in compared with the typical ensemble classifiers and ensemble classifiers of extreme learning machines, therefore demonstrating the proposed method could improve the detection accuracy of ensemble classifier.