Feature selection for image steganalysis using levy flight-based grey wolf optimization

Image steganalysis is the process of detecting the availability of hidden messages in the cover images. Therefore, it may be considered as a classification problem which categorizes an image either into a cover images or a stego image. Feature selection is one of the important phases of image steganalysis which can increase its computational efficiency and performance. In this paper, a novel levy flight-based grey wolf optimization has been introduced which is used to select the prominent features for steganalysis algorithm from a set of original features. For the same, SPAM and AlexNet have been used to generate the high dimensional features. Furthermore, the random forest classifier is used to classify the images over selected features into cover images and stego images. The experimental results show that the proposed levy flight-based grey wolf optimization shows preferable convergence precision and effectively reduces the irrelevant and redundant features while maintaining the high classification accuracy as compared to other feature selection methods.