Two-dimensional histogram expansion of wavelet coefficient for reversible data hiding

In this paper, we propose a reversible data hiding (RDH) method based on a two-dimensional wavelet coefficient histogram (2D WCH) in the wavelet domain. First, a cover image is decomposed into wavelet subbands using the invertible integer-to-integer wavelet transform (121-WT). Then, the 2D WCH is generated by counting the occurrence frequency of the wavelet coefficient pairs which denote two wavelet coefficients located in the same position in the selected two subbands where the secret message is embedded. By using the 2D WCH, the correlation between the selected subbands is more effectively utilized than the traditional ID histogram. The proposed method embed the secret message reversibly in the cover image by expanding the 2D WCH. In order to embed the secret message as efficient as possible, the expansion rule for 2D WCH is proposed. Moreover, the coefficient pair selection (CPS), which the coefficients embedding the data are selected in order to modify only the selected coefficients, is implemented before generating the 2D WCH. In the experiment, the proposed method is compared with the conventional RDH methods in terms of the capacity-distortion curve.