Reversible Data Hiding In Homomorphic Encryption Domain

With the development of technology,cloud computing and cloud security are getting more and more attention.In a cloud environment,the sharing of multimedia data would be accompanied by risks of leaks.To ensure the security of multimedia data,prevent them from being tampered and maintain the copyright of them,the administrators of the cloud environment would embed additional secret data,such as the type and capacity of the multimedia,user identity information,etc.,into the encrypted multimedia data to distinguish and effectively manage them.Driven by the needs of this scenario,researchers have begun to specialize reversible data hiding(RDH)in the encryption domain.Although RDH based on stream cipher has realized high-capacity data embedding,the security of stream cipher has a certain risk of deciphering compared to public key encryption.To overcome this problem,scientific researchers have focused on the special public key encryption system,homomorphic encryption system,using its homomorphic properties to quickly realize RDH in homomorphic encryption domain.At present,there are few RDH methods in the homomorphic encryption domain,but the potential value of this subject in cloud computing and cloud environment is huge,so this subject has much important research value.In this paper,two new schemes are proposed to study RDH in homomorphic encryption domain:1.A new high-capacity RDH scheme in homomorphic encryption domain based on preprocessing.This scheme first vacates room by preprocessing the original images,then the secret data is embedded into the room by the self-binding and the homomorphic additive of the Paillier encryption system in the encryption domain.Through the proposed scheme,the embedding rate of the embedding data can exceed 1.0 bit per pixel(bpp).By experimental verification,when the embedding rate is the same,compared with the existing schemes,the proposed scheme has better direct decryption image quality and lower distortion rate.2.A new high-capacity RDH scheme in homomorphic encryption domain by double encrypting.The scheme selects two sets of Paillier homomorphic encryption systems with different numbers bits and uses Paillier system's cipher-text randomness and self-binding properties to embed secret data into the encrypted image.Then the embedding data could be extracted in the encryption domain and the decryption domain.Although the embedding rate of this proposed scheme is not as good as the existing excellent schemes in the decryption domain,the innovation of this scheme is to ensure that the directly decrypted image is the original image,and to separate the secret data from the encrypted image.