The Use of Convolutional Neural Network on Image steganalysis: A survey

Abstract

Steganalysis is the method of recognizing the existence of concealed messages over digital multimedia. These messages are concealed using steganography techniques in digital media. Steganalysis is a challenging task with the emergence of strong Steganography algorithms. Over the past few years, steganalysis has advanced significantly due to the development of deep learning methods. In this article, we present a comprehensive review of the most recent efforts on image steganalysis in spatial and transform domains. We focused on reviewing and analyzing the most recent works that utilize convolutional neural networks in image steganalysis. Also, the technical challenges of existing approaches are discussed, along with several exciting avenues for CNN-based steganalysis