Adaptive Image Steganography Domain: A Review of the Recent Works

A steganogram is a technique for hiding a message in a cover image; the message could be an image, an audio recording, a text, or even a video. As a result of embedding, the secret message is concealed in such a way that it cannot be noticed by humans. In recent years, deep learning (DL) has become increasingly popular in various areas, including image steganography. This article focuses mainly on examining and discussing some of the current DL methods used in image steganography. For image steganography, genetic algorithm–based and general adversarial network–based methods are commonly used. It is anticipated that the findings of this study will help scholars better understand and explain several of the potential challenges in this field, as well as discuss future research directions in this area.