Research On Information Hiding In Encrypted Image For Privacy Protection

With the development of computer network technology and communication technology as well as the widespread application of cloud computing,digital multimedia resources are widely used in the Internet.How to ensure the security of the digital multimedia has become an important research topic,especially for digital image.To protect the privacy of multimedia data such as image,encryption and information hiding are two effective methods.Encryption can protect the content of an image,but the encrypted image may still be vulnerable to attacks such as tampering,piracy and forgery.Information hiding can embed secret information in the image to provide services such as storage management,copyright protection,content authentication,tamper detection and tamper recovery.However,information hiding cannot prevent an attacker from accessing the image content.Therefore,it is worthy of further study how to utilize both encryption and information hiding to protect the privacy in the applications of image storage and transmission.There are several frameworks for information hiding in encrypted image.In view of the practical application limitation of “the framework of plain image preprocessing”,this thesis only focuses on the information hiding in image after encryption.Since encryption algorithm inevitably affects the performance of information hiding,the hiding performance should be ensured under the premise of privacy protection.In particular,since compressive sensing(CS)can simultaneously complete compression and encryption with the low complexity,and the cloud computing with abundant computing resource could handle the CS reconstruction problem with high computational complexity,this thesis also introduces CS to the privacy protection application of image in cloud.According to different purposes of information hiding for privacy protection in encrypted images,this thesis mainly conducts research on storage management,tampering detection and recovery,and content authentication applications of information hiding in encrypted image.Therefore,the work of this thesis focuses on the following aspects:1 A novel reversible information hiding scheme in encrypted image based on pixel value ordered is proposed by introducing homomorphic cryptosystem.On the one hand,the innovation is that homomorphic encryption will not result in data expansion,that is to say,the volume of image is preserved after encryption.On the other hand,the proposed scheme is not only reversible but also commutative in the phase of information hiding.The extracted information and recovered image are completely consistent with the original ones.Besides,the order of information extraction and image decryption are commutative.2 On the basis of the security flaw analysis of information hiding in encrypted image based on redundant transformation,a high-capacity reversible information hiding scheme,which is capable of resisting chosen plaintext attack(CPA)and combined with chaotic system,is proposed.In the encryption process,the entropy of the original image is introduced into the generation of a parameter in chaotic system so that the encrypted image has the capability of resisting CPA.During block coding in the information hiding phase,the size information and prior information of a block are utilized to obtain a higher compression rate,which provides a higher capacity for information hiding.3 In cloud storage,a secure image tampering detection and self-recovery scheme is proposed,which is motivated by secret sharing.The image is distributed to multiple secret shares and uploaded to multiple clouds based on a permutation ordered binary system.A detection watermark and a recovery watermark are embedded in secret shares.Thus,the scheme is effective to solve the problem of tampering detection.Moreover,the tampered image can be recovered using embedded recovery watermark.In view of the local correlation of a nature image,we can improve the accuracy of the detection and the quality of restored image further.4 After analyzing the opportunity ofcombining CS encryption with cloud computing service,we propose a secure image authentication service model based on CS over cloud.A watermark image is simultaneously embedded in the encrypted image so that image authentication can be implemented while image privacy protection has been achieved.In addition,a user could use cloud computing resources to solve the problem of CS reconstruction in this model,which is also suitable for resource-constrained devices.