Article Preview
Top1 Introduction
More and more people were aware of the importance of protecting intellectual property rights. In order to protect their personal privacy and protect their legitimate rights and interests, they paid more and more attention to the protection of image property rights. Encryption technology and watermarking technology were two important means to protect multimedia digital information security and integrity. Reversible digital watermarking technology (Thodi & Rodriguez, 2007) was a new branch of digital watermarking technology. It could not only extract watermarking information correctly, but also restore the original carrier without distortion after extracting information. Therefore, it was widely used in sensitive areas such as medical images, military images and forensic images. Image encryption technology (Yun-peng et al., 2009) was to transform plaintext data into ciphertext data, which can reduce the risk of being stolen by others. How to combine encryption technology with reversible watermarking technology to improve the reliability of embedding and extracting process was a subject worthy of study.
Reversible watermarking technology mainly included spatial domain algorithm (Xiao et al., 2019)(Ishtiaq et al., 2018) and transform domain algorithm. (Nguyen et al., 2016) Spatial domain algorithm mainly used redundancy between pixels to realize watermarking embedding, and transform domain algorithm used frequency of image to embedding watermarking. Reversible watermarking technology in encryption domain belonged to a special kind, it was encrypted in space domain and transform domain. Reversible data hiding of encrypted image was widely used in medical field(Kittawi & Al-Haj, 2017)(Parah & Ahad, 2018)(Abbasi & Memon, 2018). In the aspect of algorithm, for example, Zhang (2011) used the least significant bit method to embed the watermarking after excepting or encrypting the image in the spatial domain. Hong (2012) improved the above algorithm, making full use of the pixels calculated the smoothness of each block and the pixel correlation of adjacent block boundaries. Qian (2014) aimed to encrypt a JPEG bit stream into a properly organized structure, and embedded a secret message into the encrypted bit stream by slightly modifying the JPEG stream. Peng (2019) proposed an encryption domain RDH scheme for two-dimensional vector graphics based on real reversible mapping model (Bouridah et al., 2017)(Gao & Gao,). Encryption used chaotic mode to encrypt the image. After chaotic encryption, reversible data was hidden and extracted, which had higher security. But after encryption, the redundancy between the pixels of the image was lost, this situation resulted in the difficulty of watermarking embedding and the decreased of the embedding capacity of the watermarking. Zhang improved the watermarking embedding method of encrypted image and designed a special encryption scheme to encrypt the estimation error(Zhang et al., 2014). Later, a framework of RDH-EI based on Reversible Image Transform (RIT) was proposed to meet different image quality requirements and larger embedding capacity(Zhang et al., 2016). Aiming at the problem of low embedding capacity of encrypted image(Cao et al., 2016)(Choi & Pun, 2017), different high embedding capacity methods were adopted to improve the embedding capacity.