An Efficient Steganalytic Algorithm based on Contourlet with GLCM

Steganalysis is a technique to detect the hidden embedded information in the provided data. This study proposes a novel steganalytic algorithm which distinguishes between the normal and the stego image. III level contourlet is exploited in this study. Contourlet is known for its ability to capture the intrinsic geometrical structure of an image. Here, the lowest frequency component of each level is obtained. The pixel distance is taken as 1 and the directions considered are 0, 45, 90 and 180°, respectively. Finally, Support Vector Machine (SVM) is used as the classifier to differentiate between the normal and the stego image. This steganalytic system is tested with DWT, Ridgelet, Contourlet, Curvelet, Bandelet and Shearlet. All these were tested in the aspects of first order, Run length and Gray-Level Co-occurrence Matrix (GLCM) features. Among all these, Contourlet with GLCM shows the maximum accuracy of 98.79% and has the lowest misclassification rate of 1.21 and are presented in graphs.