Fidelity preserved data hiding in encrypted images based on homomorphism and matrix embedding

Abstract

This paper proposes a data hiding scheme for high quality stego-images in the encrypted images based on a homomorphic encryption algorithm and matrix embedding method. To achieve the message embedding in the encryption domain, we first present an image encryption algorithm satisfying additive homomorphism based on the stream cipher. Then, we offer a matrix embedding method using Golay code, which has good embedding efficiency. In the proposed scheme, we pre-code the least significant bit (LSB) vectors of the cover image into the codewords before image encryption, so that we can implement the matrix embedding in the encrypted image. The experimental results show that the proposed scheme has excellent security performance. We can get a high quality stego-image after the image decryption and recover the cover image into a fidelity image while providing considerable payload.