期刊
JOURNAL OF MODERN OPTICS
卷 69, 期 15, 页码 870-886出版社
TAYLOR & FRANCIS LTD
DOI: 10.1080/09500340.2022.2093415
关键词
Image hiding; generative image hiding; convolutional neural network; feedback residual
类别
资金
- National Natural Science Foundation of China [62041106]
This paper presents an image hiding algorithm based on GFR-Net, which hides multiple color secret images in a single color carrier image. The proposed algorithm has good performance in terms of payload and security.
At present, there is always a potential threat in the process of information transmission. As a way to protect data security, image hiding has attracted extensive attention. Current image hiding algorithms have insufficient resistance to deep leaning based steganalysis algorithms and relatively low hiding capacity. This paper presents an image hiding algorithm based on a generative feedback residual network (GFR-Net), which hides multiple color secret images in a single color carrier image. First, several secret images and a carrier image were fed into the image hiding network, in which the secret images were embedded into the carrier image, resulting in an output of container image. A recovery network also based on GFR-Net was designed to reconstruct the secret images from the container. The extensive experiments for hiding normal and encrypted images show that the proposed image hiding model has a good performance in terms of payload and security.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据