Suppr超能文献

一种基于FCdDNet将大尺寸图像隐藏于小尺寸图像中的方案。

A scheme of hiding large-size image into small-size image based on FCdDNet.

作者信息

Liu Lianshan, Tang Li, Tong Shanshan, Huang Yu

机构信息

College of Computer Science and Engineering, Shandong University of Science and Technology, Qingdao, Shandong, China.

出版信息

PeerJ Comput Sci. 2024 Jun 25;10:e2140. doi: 10.7717/peerj-cs.2140. eCollection 2024.

Abstract

The hiding capacity of the current information hiding field has reached a relatively high level, which can hide two color images into one color image. In order to explore a larger hidden capacity, an information hiding scheme based on an improved FCdDNet is proposed, which can hide large-size color images into small-size color images. An improved FCdDNet network is used as the main structure shared by the hidden network and the extraction network. These two networks promote and improve each other during the confrontation training process and are used in pairs. It can be seen that the proposed scheme achieves a larger information hiding capacity, and the hidden information is four times larger than the scale of the carrier image. At the same time, the visual effect after hiding is guaranteed, and the image extracted from the hidden image also has a high degree of restoration. The scheme can be applied to image authentication, secret image transmission, and other fields.

摘要

当前信息隐藏领域的隐藏容量已达到较高水平,能够将两幅彩色图像隐藏于一幅彩色图像之中。为了探索更大的隐藏容量,提出了一种基于改进型FCdDNet的信息隐藏方案,该方案可将大尺寸彩色图像隐藏于小尺寸彩色图像之中。采用改进型FCdDNet网络作为隐藏网络和提取网络共享的主要结构。这两个网络在对抗训练过程中相互促进、相互提升,并成对使用。可以看出,所提方案实现了更大的信息隐藏容量,隐藏信息比载体图像的规模大四倍。同时,保证了隐藏后的视觉效果,从隐藏图像中提取的图像也具有高度的恢复性。该方案可应用于图像认证、秘密图像传输等领域。

https://cdn.ncbi.nlm.nih.gov/pmc/blobs/f8ff/11232572/546d16ce0cfc/peerj-cs-10-2140-g001.jpg

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验