Review of Research on Deep Learning-based Image Steganography
With the rapid development of modern information technology,multimedia image gradually becomes an important means of transmitting information,whose information security received extensive attentions.Steganography is an important technology to protect the security of information transmission,where images with rich redundancy are always selected as the transmission carrier to hide the informa-tion to realize the safe transmission of secret information.In traditional image steganography schemes,loading secret information is usually realized by modifying the pixel values of the carrier image and the impact on the carrier image is offset by adopting the distortion function,nevertheless facing the issues of poor image quality,weak anti-analysis ability,and small steganographic capacity.To address the is-sues,researchers proposed a series of advanced deep learning-based image steganography schemes.In this paper,research progresses of the traditional and the deep learning-based image steganography schemes are investigated.Then,theoretical principles of image steganography and deep learning are in-troduced,and some familiar deep learning-based image steganography models at present are deeply ana-lyzed,and a comparative analysis of the models are performed.Finally,a conclusion is presented and novel research ideas and directions are proposed.
digital imageimage steganographydeep learninginformation hiding
万方数据
维普
