Image Steganalysis Based on Spatial Domain and DWT Domain Features
Changxin Liu, C. Ou-Yang, Ming Guo, Huijuan Chen · 2010
In this paper, a new image steganalysis method was proposed based on image gradient energy and entropy features, together with Fraid's proposed wavelet subband coefficients and higher-order statistics of linear prediction error features. We get 74-dimensional features extraction from images. The support vector machines (SVM) is used to class the images. Experimental results show the method can improve the detection rate compared with Fraid's algorithm with a lower false negative rate.