Research On Authenticated Secret Image Sharing Scheme In Online Social Networks

In recent years,the abuse of multiple social network platforms and advanced algorithms such as Deep Fake has led to frequent outbreaks of personal privacy data breaches.This has prompted the public to re-examine the importance of protecting personal privacy information when sharing images on social networks.In addition,the online social network image sharing scenario itself is quite complex,and it is crucial to design personalized sharing schemes based on different scenarios and protect the content of images.Secret image sharing technology can provide an important technical means for the content protection of shared.However,traditional secret image sharing schemes cannot meet the user’s autonomous privacy control sharing requirements.In such schemes,the secret share is the loss of image availability due to noise like images.In addition,in the communication process involving multiple users in social networks,users not only require reversible recovery of shared images,but also hope that additional data embedded in the share can be reversible recovered,that is,to achieve dual reversible sharing of secret images and private data.Aiming at the problem of autonomous privacy control in sharing images in online social networks and the dual reversibility of users’ requirements for secret images and private data,this thesis proposes two research schemes for secret image sharing,with the specific content as follows:(1)Autonomous privacy control and authentication sharing schemes in online social networks.This scheme combines improved polynomial secret image sharing with visual secret sharing based on random grids,and has the advantages of lossless decryption,pixel free expansion,and simple authentication.It can be applied to user independent privacy control.In addition,using the robustness of quick response codes(QR)as a carrier for secret images to be shared on social networks avoids the lossy compression of social networks to some extent.(2)Double reversible authenticated secret image sharing scheme.This scheme uses data hiding technology in encrypted images.Firstly,it uses median edge detection(MED)prediction technology to obtain hierarchical label mapping for the bit plane of the image,and compresses it into the encrypted image.Then,combined with an improved secret image sharing scheme based on the Chinese remainder theorem,the share becomes meaningful.The experimental results show that this scheme solves the problem of user reversible data hiding,and the restored secret image is also lossless.Unlike traditional sharing schemes,this scheme ingeniously applies layered embedding operations,allowing pixels with small/large prediction errors to accommodate secret bits,thereby greatly increasing the embedding load.