Adaptive reversible data hiding scheme based on difference division interpolation

Reversible data hiding (RDH) technology is an effective means to ensure the secrecy of information transmission. In particular, the interpolation-based RDH algorithm is widely used due to its advantages of large embedding capacity, good visualization, and low computational cost. However, existing interpolation-based RDH algorithms have a common problem that the algorithm performance is greatly affected by the interpolation accuracy and image texture. This paper proposed a novel difference division interpolation (DDI) method that exploits the difference between the original pixels and predicted pixels to construct the cover image (CI) so that the pixel error is within a fixed range and independent of the image texture. Meanwhile, to avoid information aggregation, an adaptive chunking strategy (ACS) is designed to replace the traditional fixed chunking mode. The chunking mode and the number of embeddable bits for every non-reference pixel can be adaptively selected according to the embedding capacity, which enhances the visual quality of images. In addition, an out-of-bounds adjustment (OA) method is introduced to minimize the distortion of the steganographic image (SI) and solve the pixel underflow and overflow. Compared with existing algorithms, experimental results show the superiority of the proposed scheme.