Deep learning based image steganography: A review

Abstract

A review of the deep learning based image steganography techniques is presented in this paper. For completeness, the recent traditional steganography techniques are also discussed briefly. The three key parameters (security, embedding capacity, and invisibility) for measuring the quality of an image steganographic technique are described. Various steganography techniques, with emphasis on the above three key parameters, are reviewed. The steganography techniques are classified here into three main categories: Traditional, Hybrid, and fully Deep Learning. The hybrid techniques are further divided into three sub‐categories: Cover Generation, Distortion Learning, and Adversarial Embedding. The fully Deep Learning techniques, based on the nature of the input, are further divided into three sub‐categories: GAN Embedding, Embedding Less, and Category Label. The main ideas of the important deep learning based steganography techniques are described. The strong and weak features of these techniques are outlined. The results reported by researchers on benchmark data sets CelebA, Bossbase, PASCAL‐VOC12, CIFAR‐100, ImageNet, and USC‐SIPI are used to evaluate the performance of various steganography techniques. Analysis of the results shows that there is scope for new suitable deep learning architectures that can improve the capacity and invisibility of image steganography. This article is categorized under: Technologies > Computational Intelligence Technologies > Machine Learning Technologies > Artificial Intelligence

Enhancing steganography capacity through multi-stage generator model in generative adversarial network based image concealment

“…The three main categories of steganography techniques are traditional, hybrid, and fully deep learning. 3 There are three subcategories of hybrid approaches: adversarial embedding, distortion learning, and cover generation. Based on the type of input, the fully deep learning techniques are further divided into three subcategories: GAN embedding, embedding less, and category label.…”

“…The three main categories of steganography techniques are traditional, hybrid, and fully deep learning 3 . There are three subcategories of hybrid approaches: adversarial embedding, distortion learning, and cover generation.…”

“…The steganography techniques are primarily categorized into three groups: traditional, hybrid, and fully deep learning 3 . Traditional methods for encoding and decoding secret information require rules created by humans.…”

“…Traditional methods for encoding and decoding secret information require rules created by humans. Three performance measures are used to assess the steganographic image’s quality: capacity, security, and invisibility or imperceptibility 3 , 4 . To increase security, capacity, and invisibility, many steganography techniques have been put forth in the literature 5 …”

“…These steganographic techniques do not directly relate the distortion function to undetectability; instead, it is heuristically built 12 . Deep learning has been explored by researchers in various applications, such as image classification, 14 , 15 image steganography and steganalysis, 3 , 16 , 17 and other applications 18 . Given the recent quick advancement of deep learning, steganalysis methods based on deep learning, such as ATS, 19 XuNet, 13 YeNet, 20 SRM, 21 and Yedroudj-Net, 22 have all demonstrated results that are satisfactory.…”