参数化二叉树编码实现加密图像可逆数据隐藏

为了提高数据嵌入容量，提出了一种基于参数化二叉树编码的加密图像可逆数据隐藏新方法。该方法的系统由图像所有者、数据嵌入者和授权接收者三方组成，其中，图像所有者采用特殊的图像块随机调制和置乱对原始图像加密并上传；由于加密操作很好地保留了图像块的像素相关性，数据嵌入者可在加密域充分利用图像块相关性，通过自适应选择参考像素来预测同一图像块的其余像素得到预测差值，再使用参数化二叉树编码进行无损压缩生成嵌入空间，实现大容量数据嵌入；授权接收者可正确提取嵌入数据或无失真恢复原图像。仿真实验结果表明：所提方法提高了嵌入容量，且可用于医学、云服务、军事等众多领域。

Reversible Data Hiding in Encrypted Image Using Parametric Binary Tree Labeling

To achieve high embedding capacity, an improved reversible data hiding method based on parametric binary tree labeling is proposed for encrypted images. The proposed method includes the following three independent stages:image owner, data hider and authorized receivers. In the first stage, image owner encrypts the original image in a special way of modulating randomly and scrambling image blocks but reserving the pixels' correlation within image blocks. After the encrypted image is uploaded onto cloud servers, data hider first chooses a reference pixel of each image block by adoptive way, and predicts the rest pixels according to the correlation information. Then the data hider can embed large amounts of additional data into the encrypted image by using parametric binary tree labeling in the second stage. In the last stage, authorized receivers can extract the embedded data, or recover the original image losslessly. Experimental results show that the improved method can not only achieve high embedding capacity, but also be applicable in medical, cloud services, military and other fields.