具有超分辨率行为伪装效果的可逆图像隐藏  

Reversible image hiding with super-resolution behaviorcamouflage effect

作　　者：贾孟霖 杨杨[1,2] 孙冬 Jia Menglin;Yang Yang;Sun Dong(School of Electronics and Information Engineering,Anhui University,Hefei 230601,China;Institute of Artificial Intelligence,Hefei Comprehensive National Science Center,Hefei 230026,China;School of Electrical Engineering and Automation,Anhui University,Hefei 230601,China)

机构地区：[1]安徽大学电子信息工程学院,合肥230601 [2]合肥综合性国家科学中心人工智能研究院,合肥230026 [3]安徽大学电气工程与自动化学院,合肥230601

出　　处：《中国图象图形学报》2024年第2期382-394,共13页Journal of Image and Graphics

基　　金：国家自然科学基金项目(62272003);安徽省高等学校自然科学基金项目(KJ2021A0016);教育部产学研协同育人项目(202102246002)。

摘　　要：目的 图像隐藏已成为计算机视觉领域的一个重要课题,其目的是以难以察觉的方式将秘密图像隐藏在载体图像中,同时要求接收端能够恢复秘密图像。尽管该技术发展迅速,但目前的图像隐藏技术大多是从内容层面进行伪装,追求载密图像与载体图像的不可区分性。其实,图像隐藏的本质是对行为安全的追求,因此不仅可以在内容层面进行伪装,还可以在行为层面进行伪装。方法 本文从行为安全的角度出发,提出了一种基于超分辨率行为伪装的可逆图像隐藏方法。与传统的图像隐藏技术不同,本文首先将秘密图像可逆地隐藏到载体图像中,生成载密图像,然后通过可逆的超分辨率处理创建与普通超分辨率图像处理操作无法区分的伪装图像。最后,允许接收方从伪装图像中恢复秘密图像和载体图像。结果 在图像隐藏和超分辨率两个任务中,本文方法均取得了优异的结果。在相同的数据集下,测试结果显示恢复秘密图像的峰值信噪比(peak signal-to-noise ratio, PSNR)值达到47+dB,较对比方法提升了2%以上,结构相似度(structure similarity index measure, SSIM)值也达到0.99+,超分辨率图像与Bicubic、SRCNN(super-resolution convolutional neural network)方法的结果相比,峰值信噪比(PSNR)提升了2+dB,感知指数(perceptual index, PI)值降低了2.02+。结论 本文提出的图像隐藏框架利用可逆超分辨率处理操作实现了行为安全角度的图像隐藏,在容量、安全性和精度上都具有优势。Objective Image hiding has recently become a hotspot in the computer vision field.It aims to hide the secretimage in a cover image imperceptibly and recover the secret image,preferably at the receiver.Traditional image hidingmethods often adjust the cover image’s pixel value in the spatial domain or modify the cover image’s frequency coefficientsto hide the secret information.These methods hide secret images through handcrafted feature information.Thus,these secret images can be detected easily by existing detection techniques.These methods have weak security and lack signifi⁃cant capabilities for hiding information in images.Therefore,they fall short of meeting the demands of large-capacity imagehiding tasks.Image hiding methods based on deep learning have been quickly developed with the advancements in convolu⁃tional neural networks.These deep learning methods seek to achieve a high level of capacity,invisibility,and recoveryaccuracy.However,the existing image hiding techniques can be easily detected by deep learning analysis methods becauseof the rapid development of steganalysis.Handcrafted image hiding methods and image hiding methods based on deeplearning are camouflaged from the content level to pursue the indistinguishability of the marked image and the cover image.The essence of image hiding is the pursuit of behavioral security;that is,the pursuit of hiding secret information is insepa⁃rable from the behavior of normal users to achieve good detection resistance.Therefore,we can camouflage at the contentlevel and disguise at the behavior level.We innovatively use super-resolution,a common image processing technology,asour behavior camouflage means to realize image hiding from the behavior security perspective.Method In general,tradi⁃tional image hiding techniques tend to prioritize the indistinguishability of the cover image and the secret image at the con⁃tent level.However,we aim to achieve image hiding from a behavioral security perspective in our study.In particular,weaim to make stegan

关 键 词：图像隐藏 可逆神经网络(INN) 行为安全 可逆行为伪装 超分辨率 深度学习 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]