Position-Aware Guided Hiding Data Scheme with Reversibility and Adaptivity for Dual Images

Abstract

Reversible data hiding (RDH) in dual images is a technique that shares secret messages into two similar shadow images, while the secret messages and the cover image can be restored only when those two shadows are gathered simultaneously. In this paper, a novel turtle shell-based RDH hiding scheme based on the symmetric property is presented in order to increase the embedding capacity and maintain good visual quality in dual images under the guidance of position-aware. First, we classify each pixel pair into one of four types according to their locations and then determine a sunflower area centered around it in order to construct the combination of positions and the embedding table. Using the embedding table, the secret messages are concealed into a cover image by generating two shadow images. At the decoder’s side, the complete restoration of the secret messages and the cover image can be accomplished by identifying the position relationship between the two stego-pixel pairs. The experimental results confirmed that the proposed position-aware guided RDH scheme is superior to some of the relevant works on the aspects of embedding capacity or image quality. In addition, the proposed scheme provides a secure communication that can effectively resist attacks on the pixel value difference histogram, relative entropy, and regular singular analysis.