Article Preview
Top1. Introduction
Data hiding technique (Ma, 2016 & Zhang, 2016), which is a rising and effective tool used for multimedia processing and other related fields (Du, 2016; Gopinath, 2013; Liu, 2014; Petrlic, 2013; Radwan, 2009; Thabet, 2014; Zhu, 2013) has attracted more and more attentions. Reversible data hiding (RDH) is a novel technique that allows the original image to be recovered losslessly after the embedded data are extracted. With this property, RDH is widely used in many specific fields including image processing (Liu, 2016; Uchida, 2017; Wang, 2016), web security (Kuniyasu, 2018), etc. With the development of cloud computing and web services, many cover images are transformed into cipher image before uploading, for the purpose of security protection. Reversible data hiding for encrypted image (RDHEI) has become the research hotspot recently, and has many important applications in web security (Di, 2017; Qian, 2016). Specifically, some medical images in the cloud have been encrypted in former, due to the purpose of privacy protecting. Thus, someone in the system need embed some authentication or management message for convenience. However, both the original image and the embedded message need to be recovered.
Existing RDH methods in plain domain such as the difference expansion method (Tian, 2003), histogram shifting method (Dragoi & Coltuc, 2014) and lossless compression method (Jarali & Rao, 2013) are unsuitable for the RDH in the encrypted domain. Zhang (2011) firstly proposes the RDH method in the encrypted domain using flipping the pixel values. In this method, the additional data is embedded into images which are encrypted with stream cipher. The cover image is recovered by the correlation among the pixels. Hong et al. (Hong, Chen, & Wu, 2012) proposed the improved method of Zhang’s algorithm, but this algorithm did not achieve better performance when the size of the block is small. Many RDHEI methods (Cao, 2016; Ma, 2013; Nyuyen, 2016; Xu, 2016; Zhang, 2014; Zhou, 2016) have improved the embedding performance in recent years.