Abstract: Medical image facing malicious tampering, information thefting and leaking, this paper proposes a hyper-chaos based zero watermarking scheme for image authentication and tamper detection. Firstly, original carrier image’s LSB plane should be initialized to zero. Secondly it is divided into non-overlapping blocks and calculating the mean value of each block, using the size relationship between each pixel of block and mean value of each block to construct the characteristic matrix. Finally, Arnold scrambled characteristics matrix and hyper-chaotic encrypted binary watermark are executed for XOR operation to construct a zero-watermark. In addition, if the image has been tampered, tampered area can be detected and localized accuracy by calculating the difference image. The experimental results show that the scheme not only has good security, but also achieves superior tamper detection and localization.

Key words: hyper-chaos encryption, tamper detection, zero watermarking, medical image

摘要： 摘  要：针对医学图像面临恶意篡改、信息窃取和泄露等安全问题，提出了一种基于超混沌的医学图像篡改定位零水印算法。首先将原始载体图像的最低有效位LSB（Least Singificant Bit）置零，然后将其分成互不重叠的子块，计算每块的均值，利用块中每个像素值与块均值的大小关系构造特征矩阵，最后利用Arnold置乱后的特征矩阵与超混沌加密的二值水印进行异或运算来构造零水印；另外，如果图像遭到篡改，通过计算差值图像，则能精确定位篡改位置及篡改形状。实验结果表明，该方案不仅具有较好的安全性，而且达到良好的篡改检测和定位效果。

关键词: 超混沌, 篡改检测, 零水印, 医学图像