加密医学图像的可逆信息隐藏算法研究重庆大学硕士学位论文（学术学位）学生姓名：程惠惠 指导教师：郑洪英副教授 专业：计算机软件与理论 学科门类：工学重庆大学计算机学院二 O 一五年四月Study on Reversible Data Hiding Algorithm for Encrypted Medical ImageA Thesis Submitted to Chongqing University in Partial Fulfillment of the Requirement for theMasters Degree of EngineeringBy Cheng HuihuiSupervised by Associate Prof. Zheng Hongying Specialty：Computer Software and TheoryCollege of Computer Science of Chongqing University, Chongqing, China April, 2015摘要医学影像传输标准 DICOM(Digital Imaging and Communications in Medical)为 数字医学图像在通信网络上的显示、传输和存储，提供了标准的格式。DICOM 标 准简化了医学影像信息间的交换，进一步推动了远程医疗系统的数字化进程。随 着计算机科学、通信技术的不断发展以及 DICOM 标准的普及和远程医疗的实施， 医学图像的保密与认证等安全性保护显得越来越重要。近年来，为了实现信息的安全传输，人们提出了各种类型的安全技术，密码 学和信息隐藏技术是其中最重要的两个分支。密码学是为了隐藏明文信息的内容， 而信息隐藏技术则是为了隐藏信息的存在性。而在现实应用中通常需要将加密和 信息隐藏技术结合起来以提高数据的安全性，结合以后不但能提供机密性保护， 而且能够提供完整性保护。因此，随着远程医疗技术的发展，基于医学图像加密 域的可逆信息隐藏算法的研究具有十分重要的研究意义和实用价值。本文所做工作主要体现在以下三个方面：对数字图像现有的压缩、加密和信息隐藏算法进行深入研究，试图设计出 适用于加密医学图像的可逆信息隐藏算法。为了实现完全可逆信息隐藏算法和提高嵌入容量，结合边缘检测、行程编 码(RLC)、混沌加密和 LSB 替换算法，提出了一种基于局部压缩的密文域可逆信 息隐藏算法。首先使用 SUSAN 边缘检测算法检测边界信息，用 RLC 压缩边界部 分并重构图像，然后使用混沌加密算法加密图像，最后通过 LSB 替换的方法将水 印信息嵌入加密后的图像中。根据接收方拥有的密钥类型的不同，有三种不同的 操作情况。Matlab 仿真实验表明，该算法具有较高的嵌入容量，且解密和水印提 取的操作顺序可以根据需要动态选择。结合 RLC、置乱加密和直方图平移算法，提出了一种基于整体压缩的密文 域可逆信息隐藏算法。首先将图像做位平面分解操作，使用 RLC 压缩最高的四个 位平面并重组图像，然后通过由 Logistic 混沌映射产生的序列对图像作置乱加密， 最后利用直方图平移方法将水印嵌入到密文图像中。该算法适用于各种类型的数 字图像（如无零点图像），将峰值点和零值点存储在压缩后预留的空间中，无需 另外传输，且接收端的操作灵活。最后总结了全文，并对未来密文域信息隐藏的研究工作进行了展望。关键词 ： 医 学 图 像 ， 边 缘 检 测 ， 行 程 编 码 ， 图 像 加 密 ， 可 逆 信 息 隐 藏ABSTRACTThe Digital Imaging and Communications in Medical (DICOM) is the standard for formatting the digital medical images displaying, exchanging and storing in communication network. DICOM standard simplifies the exchange of medical image information, and further promotes the digital process of telemedicine. With the continuous development of computer science and communication technology, the popularity of DICOM standard and the implementation of telemedicine, confidentiality and authentication protection of medical image are becoming more and more important.Recently, various types of security technology are proposed to improve the security of the information transmission, cryptography and information hiding technology are the two most important branches. Cryptography is to hide the content of plain information, and the information hiding technology makes the existence of information invisible to the illegal users. In practice, the combination of encryption and data hiding technology usually are required to improve the data security, which provides not only confidentiality protection, but also integrity protection. Therefore, with the development of telemedicine technology, research on reversible data hiding algorithm based on encrypted medical image is important, significance and practical value.The works of this article are as follows:Firstly, research and analysis on the existing algorithms of medical image compression, cryptography and data hiding, so as to design the reversible data hiding algorithm based on encrypted medical images.Secondly, in order to implement reversibility of data hiding technology and improve the information embedding rate, a reversible data hiding algorithm for encrypted images based on a partial compression is proposed by using of edge detection, run length coding, chaotic encryption and LSB substitution. At first, the edge information of an image is detected by SUSAN algorithm, and the image is reconstructed after compressing boundary portion by means of run length coding (RLC). Then the image is encrypted using chaotic algorithm and the LSB replacement method is used to embed the watermark in the encrypted image. In the receiving side, there are three operations to choose according to the obtained keys for the receiver. The simulations on MATLAB indicate that the algorithm has a high embedding capacity, and the operating sequence of decryption and watermark extraction is flexible.Thirdly, combination of RLC, scrambling encryption and histogram modification, a reversible data hiding algorithm for encrypted images based on an overall compression is proposed. At first, the image is decomposed into 8 independent bit-planes, the higher four bit-plans are compressed using RLC and the image is reconstructed. Then the image is scrambled according to the sequence which generated by Logistic chaotic map. Finally, information is embedded into the encrypted image by using histogram modification. The algorithm is suitable for all kinds of digital images (for example the images which cannot find zero point), the peak point and the zero point is stored in the reserved room, so there is no need to send the peak point and the zero point, and the operation sequence at the receiving end is flexible.Finally, we summarize the article and indicate the future research of data hiding for enc