A high capacity reversible data hiding scheme based on right-left shift

Abstract

Abstract Recently, data hiding scheme developed rapidly. But its permanent distortion does not satisfy some sensitive applications. Accordingly, reversible data hiding (RDH) has developed quickly for years. Histogram shifting (HS) technique is one of the high-efficiency RDH schemes. In this paper, based on the characteristics of rectangle-prediction-error distribution, i.e., bin 0 is the highest bin, the other peak bins are distributed on the left and right sides of the highest bin in a nearly symmetrical way, and many zero bins locate on both sides of the peak bins. We proposed a high capacity RDH scheme based on right-left shift. The peak bins are first shifted towards right direction to leave some locations for embedding secret message bits, then, they are shifted towards left direction leave more locations for more data embedding, which can reduce the distortion due to right shift. By analyzing the necessary condition for the predictive error, the issue of overflow/underflow can be easily solved without adding side information in the embedding data. Experimental results on different images illustrate the proposed scheme is of high embedding capacity (EC) and visual quality on different images, meanwhile, the theory demonstrates the performance of the proposed method outperforms other RDH methods.