基于预测误差位平面压缩的密文图像可逆信息隐藏

To further improve the performance of reversible data hiding in encrypted image, an algorithm for lossless compression of the prediction error bit-plane using joint encoding was proposed, which could make full use of image redundancy and reserve more embedding room. Firstly, the image owner calculated the prediction error of the image and divided the prediction error bit-plane into non-overlapping blocks of the same size. Then, the prediction error bit-plane was rearranged according to blocks and the rearranged bitstream was compressed by run-length encoding and Huffman encoding to reserve room. The data hider embedded information in the reserved room of the encrypted image. At the receiving end, the legitimate receiver extracted information and recovered images losslessly and separately. Experimental results show that the proposed algorithm makes full use of the bit-plane distribution characteristics and achieves higher embedding performance. The average embedding rates in BOSSbase and BOWS-2 datasets reach 3.763 bpp and 3.642 bpp, which are at least 0.081 bpp and 0.058 bpp higher than the state-of-the-art algorithms.