Study Of Reversible Data Hiding Scheme In Homomorphic Encrypted Domain

As cloud computing develops rapidly, massive data are uploaded and stored in the cloud or remote server after encryption. Since the security threats exist, a big challenge lies in the security protection and integrity authentication of the encrypted data. Besides, the management of massive ciphertexts remains a grave problem. Reversible data hiding in encrypted image emerges as these situations require, which combines cryptography with reversible data hiding technology. Reversible data hiding in encrypted image, which can embed additional data directly into encrypted image and recover original image perfectly after encryption and data extraction, is applied to achieve efficient management and security protection of the encrypted image. Since the public key cryptosystem allows algebraic operations of cipertexts, reversible data hiding in homomorphic encrypted domain has become a hot spot in the study of information security. In this paper, two new reversible data hiding scheme in encrypted domain is presented:1) A novel reversible data hiding scheme in homomorphic encrypted domain based on reserving room before encryption is proposed. The proposed method reserves target pixels by embedding them into other part of image with standard reversible data hiding algorithm. As a result, data hiding procedure is accomplished by replacing the ciphertexts of target pixels with an encrypted weighted sum of additional data. On the receiver side, the hidden data can be extracted by employing the value match method. Experimental results demonstrate that the computation complexity is extremely low and original image can be recovered losslessly after decryption and data extraction.2) A novel reversible data hiding scheme in homomorphic encrypted domain based on mirror ciphertext pair is proposed. Method of reserving room before encryption is utilized. In each target pixel pair, the mirror pixel and least significant bit of host pixel are reserved before encryption. Content owner constructs the mirror ciphertext pair after encryption. As a bonus, additional data can be embedded into the reserved room by implementing homomorphic multiplication. Receiver can derive the corresponding ciphertexts of additional data with modular multiplicative inverse method and extract the original hidden data with value match method. The computation complexity decreases to O(k), where k is the embedding capacity. Data hiding and data extraction in encrypted domain can be accomplished promptly. Besides, the proposed scheme has a good performance on PSNR since the most significant bits are unchanged. Most importantly, additional data are embedded into the bit-planes of corresponding plaintexts in encrypted domain and capsuled with the homomorphic and probabilistic properties of cryptosystem. The hidden data in the proposed method has a higher security.