A novel pixel value ordering reversible data hiding based on dual-image

Abstract

Recently, reversible data hiding (RDH) based on pixel value order (PVO) has been an important research direction in the field of information security due to its high-fidelity characteristics. However, PVO-based schemes have low embedding capacities (EC) because the pixels whose prediction-error greater than 1 or less than -1 on the histogram just shifted but not selected for embedding data. Dual-image techniques have often been used in recent years. In this paper, we use dual-image to design a new shift strategy that allows secret data to be embedded in those pixels only shifted in PVO-based schemes previously to improve the EC. The number of secret message bits for each embedding is fixed in most dual-image RDH methods, which leads to image redundancy not be fully exploited. In the proposed method, the number of embedded bits changes dynamically according to the prediction-error so that the visual quality of the stego images is better. Experimental results verify that the proposed method is superior to some other state-of-the-art PVO-based and dual-image RDH methods.