A Markov-based reversible data hiding method based on histogram shifting

Abstract

Applying information theory, this work considers an image as a stream of symbols emitted by a Markov information source. With the Markov model, a reversible data-hiding scheme based on the histogram modification technique is proposed to provide an efficient tradeoff between hiding capacity and quality of a marked image by changing the order of the Markov model. The larger the order is, the higher the capacity is but the lower the quality is, and vice versa. The experimental results show that the proposed scheme yields not only much larger hiding capacity but also smaller image distortion than other reversible data-hiding schemes reported in the literature. This work also proposes two feasible approaches to reduce the overhead yielded during the data embedding.