Cyber Security and Network Security

6Insights Into Deep Steganography: A Study of Steganography Automation and Trends

R. Gurunath¹, Debabrata Samanta^1* and Digvijay Pandey²

¹Department of Computer Science, CHRIST (Deemed to be University), Bangalore, India

²Department of Electronics Engineering, Institute of Engineering and Technology, Dr. A.P.J. Abdul Kalam Technical University, Lucknow, India

Abstract

Recurrent neural networks (RNNs) are built on the foundation of feed forward networks. The greatest comparison for RNN is simple writing analysis, where the prediction of the next word is always dependent on prior knowledge of the sentence’s contents. RNN is a type of artificial neural network that mimics the human neuron network and is used to recognize a series of data and then analyze the results to anticipate the conclusion. The LSTM is a kind of RNN that comprises of a stack of layers containing neurons. This article also discusses the problems that each technology faces, as well as potential solutions. To reduce losses, optimization algorithms change the characteristics of neural networks, such as weights and learning rates. One of the sections provides optimization algorithms in neural networks. A section devoted to some of the most recent extensive research on steganography and neural network combinations. Finally, we present an analysis of existing research on present study for the previous 5 years (2016 to 2020).

Keywords: Stenography, convolutional, CNN, NLP, RNN, LSTM

6.1 Introduction

In text steganography, ...