Reversible data hiding based on global adaptive pairing and optimal 2D mapping set

Two-dimensional prediction error expansion(2D-PEE) can effectively improve performance of reversible data hiding (RDH) because it can make full use of correlation of predicting errors. Most of the existing 2D-PEE is based on a fixed pairing pattern of adjacent positions. The fixed 2D mapping through experience obtained has the same mapping rules for different texture images, this will limit embedding performance. To address the problem, this paper proposes an RDH scheme based on global adaptive pairing and optimal 2D mapping set. The global adaptive pairing method is adopted to directly pair the ordered prediction error sequence. Since the prediction error pair is not constrained by position, the two-dimensional prediction error histogram (2D-PEH) is sharp. The single 2D-PEH is divided into multiple 2D-PEHs, and the 2D mapping of each sub-2D-PEH is adaptive determined by the dynamic programming method. Thereby, the optimal 2D mapping set is self-adaptive obtain according to the texture feature of the image. Experimental results show that the proposed scheme outperforms other state-of-the-art schemes. The PSNR of the image Lena is reaches 60.84 dB for an embedding capacity of 10,000 bits.