Reversible data hiding in encrypted image with local-correlation-based classification and adaptive encoding strategy

• A local-correlation-based classification mechanism is proposed such that these blocks of an image can be classified into different types. • The data hider adopts multi-layer processing to fully exploit the local correlation in each block. • For those blocks whose correlation is not well preserved, we propose a special processing method to exploit redundant space In recent years, reversible data hiding in encrypted images has attracted more and more attention due to its wide application in secret information transmission and image content protection. In this paper, we propose a reversible data hiding scheme for encrypted image based on classification and multi-layer processing. The image owner utilizes a block-level encryption method, including block-level bitwise exclusive-or operation and block permutation, to obtain encrypted image such that the correlation of pixels in most blocks is preserved. Therefore, there provably exist identical consecutive high-order bit planes (ICHOBPs) from the most significant bit (MSB) where the number of ICHOBPs is defined as local correlation. Further, a local-correlation-based classification mechanism is proposed such that these blocks can be classified into different types. The data hider adopts multi-layer processing to fully exploit the local correlation in each block. First, ICHOBPs of each block are utilized as redundant space for data embedding. For those blocks with a small number of ICHOBPs, an adaptive coding strategy is proposed to improve embedding capacity. Especially for those blocks whose correlation is not well preserved, we propose a special processing method to exploit redundant space. Experimental results show that our method can effectively improve the embedding capacity than other similar ones with keeping high quality in security analysis.