Research On Compressive Sensing-based Information Hiding In Cipher Image

The rapid development of computer network and multimedia information processing technology has brought great convenience and efficiency to our daily life.Meanwhile,more and more digital information stored in the open network environment is facing the unprecedented risks of illegal access or malicious tempering.Therefore,privacy protection and security management of digital information have become particularly prominent.Cryptography is the main technique to realize data privacy protection.Information hiding is the important method to manage and protect the host data and achieve secret information transmission securely.Combining the two technologies can achieve privacy-preserved management of the host data by embedding information into the encrypted domain,and this research topic is hot.At the same time,compressive sensing,as a widely concerned new technology,can simultaneously realize sampling and compression,which is especially suitable for the processing of digital images with high redundancy.Considering that the compressive sensing technology can also realize encryption by combining some cryptographic techniques,it is of great significance to research information hiding in encrypted image based on compressive sensing technology.By studying the relevant literatures,two novel schemes about information hiding in encrypted image based on compressive sensing are proposed in this paper.(1)Privacy-preserved information hiding in encrypted image by using compressive sensing and FCM clustering.On the sender side,the content owner preprocesses the original image first,that is,the original image is divided into several highly correlated classes by FCM clustering algorithm.Then,by setting the appropriate threshold,all the classes are further classified into two parts.After that,one part is encrypted with traditional stream cipher,and the other is compressed and encrypted simultaneously using the compressive sensing technology.This can not only protect the security of original information,but also provide great convenience for information management by the untrusted third-party.On the information hider side,one can directly embed additional information into the space vacated in the first phase.At the receiver side,the legal user can achieve information extraction and image decryption respectively according to the information hiding key and decryption key respectively,they are completely exchangeable and separable.Experimental results show that the scheme can not only guarantee the security of the carrier image,but also achieve high embedding capacity.(2)High fidelity preserved information hiding in encrypted image based on homomorphism and compressive sensing.At the sender side,the content owner uses a homomorphic encryption scheme to preserve the correlation between adjacent pixels in the original image.At the information hider side,one can directly process the signal in cipher domain according to the homomorphism of encryption system.First,the cipher image is classified into smooth set and texture set,which is exactly equal to the classification result in the plaintext domain.Then,using compressive sensing technology,only LSBs layer of smooth set are compressed to create space to accommodate additional information,and the high fidelity of the carrier image is preserved.At the receiver side,according to the information hiding key,the user can extract additional data precisely;by the decryption key,the user can obtain a directly decrypted image with high visual quality;and with both information hiding key and decryption key,not only the additional data can be extracted accurately,but also the carrier image can be recovered completely.Vast simulation results show that this method can achieve considerable embedding capacity and excellent visual quality of the host image.