Deep Learning in steganography and steganalysis from 2015 to 2018

Abstract

For almost 10 years, the detection of a hidden message in an image has been mainly carried out by the computation of Rich Models (RM), followed by classification using an Ensemble Classifier (EC). In 2015, the first study using a convolutional neural network (CNN) obtained the first results of steganalysis by Deep Learning approaching the performances of the two-step approach (EC + RM). Between 2015-2018, numerous publications have shown that it is possible to obtain improved performances, notably in spatial steganalysis, JPEG steganalysis, Selection-Channel-Aware steganalysis, and in quantitative steganalysis. This chapter deals with deep learning in steganalysis from the point of view of current methods, by presenting different neural networks from the period 2015-2018, that have been evaluated with a methodology specific to the discipline of steganalysis. The chapter is not intended to repeat the basic concepts of machine learning or deep learning. So, we will present the structure of a deep neural network, in a generic way and present the networks proposed in existing literature for the different scenarios of steganalysis, and finally, we will discuss steganography by deep learning.

Image steganalysis based on model compression

“…[10]. The automatic network architecture search technique proposed in the literature [11] has shown significant results in reducing model dimensionality and maintaining performance. Also, model pruning is an effective tool and is easy to operate and implement [12].…”

Search for Image Steganographic Policy With Adversary and Auxiliary Constrained Distance Measure

“…The adversarial game is the essential nature between steganography and steganalysis. Due to the good performance of deep learning in steganalysis [15] and many other tasks, it is rational to incorporate deep neural networks into steganography to defeat these state-ofthe-art steganalysis methods. There are mainly two strategies, i.e., learning the steganographic scheme based on Generative adversarial networks(GANs) [16]- [20] or fooling a pretrained CNN-based steganalyser by the adversarial attack [21]- [23].…”

Hiding Data Hiding

“…Most of the advanced methods use deep neural networks (DNNs) to learn the modification probabilities of cover elements so that a cost function can be determined, allowing us to embed secret data with the minimized distortion. A survey can be found in [5]. It is true that these arts have moved data hiding ahead rapidly.…”