基于预测误差分类置乱的图像加密域可逆信息隐藏

Reversible Data hiding in encrypted image based on prediction error and classification scrambling

为提高XU算法的隐藏容量和加密图像的安全性,提出了一种基于预测误差分类置乱的加密域可逆信息隐藏算法。首先根据XU的预测方法得到非采到样像素的预测误差,根据预测误差的取值范围将所有非采样像素分为可变像素和不变像素两类。对采样像素、可变像素和不变像素分类置乱后再异或加密生成加密图像,以提高加密图像的安全性。可变像素的低2位用于保存图像信息,剩余的6个高有效位均可用于隐藏信息,实现了一个可变像素嵌入多比特附加信息,提高了嵌入容量。实验结果表明,本算法的加密方法能有效抵抗COA唯密文攻击,且平滑图像的隐藏容量可以超过2比特每像素。

To improve the hidden capacity and the security of encrypted images of XU′s algorithm,a reversible data hiding algorithm based on prediction error and classification scrambling is proposed.Firstly,the prediction error of the non-sampled pixels is obtained by the prediction method of XU.Secondly,according to the range of prediction error,all non-sampling pixels are divided into two categories: variable pixels and invariant ones.The sampled pixels,the variable pixels,and the invariant pixels are respectively scrambled and then encrypted with stream cipher to generate the encrypted image.The bit-XOR with pixel-scrambling combination improves the security of the encrypted image.The two least significant bits of the variable pixels are used to store image information,and the six most significant bits of it are used to hide information.The embedding capacity is greatly improved since more than one bit are embedded in one pixel.Experimental results show that the encryption method of the proposed algorithm can effectively resist cipher-only attack(COA),and the hidden capacity of the smoothed images can exceed two bits per pixel.In general,the proposed algorithm has higher security and embedding capacity.