Recent advances for hiding content in image Steganography

An investigation into deep learning approaches for image steganography

Abstract

This paper's primary objective is to investigate and describe the several deep learningapproaches currently used for picture steganography. Traditional approaches, The most common deeplearning approaches for photo steganography are convolutional neural network (CNN) based methodsand generative adversarial network (GAN) based methods. The authors of this work set out to aid theirfellow researchers by gathering pertinent data on the most recent developments, difficulties, andpotential future directions in this area. The pictures to be concealed are embedded by first transformingthe cover image to luminance and chrominance components. An excellent example of the class ofpermutation-based algorithms that may better survive channel degradations is the chaotic Baker map,which is used to encrypt the secret pictures. In this study, an OFDM system with channel equalizationwas used for wireless communication.