Presenting a New Method of Image Steganalysis Based on MLP Neural Network

The ever-increasing development of telecommunications has made secure transmission one of the most important issues today. Since there is a high hiding capacity in the image, the use of image encryption is much more common than other methods of encryption. This article uses the covert imaging technique with the wavelet transform method, and the results show that this method has high resistance. For the analysis of hidden images, an algorithmic wavelet transform method using matrix features (GLCM) and co-occurrence vectors (DCL) is presented. After checking these values in the original and cover images, the different features between these images are extracted and used to train the multilayer neural network (MLP). The classification stage has been performed using the layers of this neural network and the proposed algorithm has been tested for a database of 200 standard images (Casia-Iris). The detection accuracy of 90% of the hidden images in the proposed method shows the superiority of this hidden mining method over other methods.