一种新型的基于CS-MRI的图像信息加密方法  

Novel method of image encryption based on CS-MRI

作　　者：简献忠[1,2] 刘钢 Jian Xianzhong;Liu Gang(School of Optical-Electrical&Computer Engineering,University of Shanghai for Science&Technology,Shanghai 200093,China;Shanghai Key Laboratory of Modern Optical System,Shanghai 200093,China)

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093 [2]上海市现代光学系统重点实验室,上海200093

出　　处：《计算机应用研究》2018年第11期3428-3431,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(41075019)

摘　　要：针对现有基于压缩感知加密方法对图像数据降维程度不足的问题,提出一种基于CS-MRI的改进加密方法。在发送端,先将图像变换到K-Space,利用随机二相编码掩模作为采样矩阵进行图像欠采样,作为第一次加密。去除欠采样得到数据中的填0项并进行随机置乱排序,完成数据的压缩与第二次加密。最后对降维后的数据使用Arnold变换进行三次加密,再将采样到的直流分量分布保存于待传输数据。在接收端,采用阈值法解出直流分量备用,在完成数据解密后使用Split-Bregman算法进行图像重构。加密后的数据降维程度与重构图像质量理想,只保留30%的数据作传输,峰值信噪比为34. 9 dB,且能够很好地抵抗噪声、裁剪、旋转攻击,鲁棒性强。实验表明,与现有CS方法相比,该方法可有效降低数据维度,且系统具备良好的安全性。For the problem of reducing dimension in image encryption based on common compressive sensing,this paper proposed a novel method improved on CS-MRI.In the sending end,firstly it transformed the image into K-Space and undersampled K-Space data with a random two phase coding mask as the first encryption.Then,it removed the fill-0-item from the undersampled data and put the remaining data to a new location as the second encryption.After getting the data which were compressed and encrypted by last two process,it encrypted this data by block Arnold trans again as the third encryption.Finally,it saved an important data called DC component into the thrice-encrypted data by a scattered way.In the receiving end,it used the threshold method to solve the DC component from received data as a backup.Then it decrypted the data by three public keys and reconstructed the original image by Split-Bregman algorithm.It had good effects in both data dimension reducing and image reconstruction,only 30%data were reserved and the PSNR could increase to 34.9 dB.In the same time,it was robust against attacking.Experiment shows that the proposed scheme gets a better dimension reducing performance compared to common CS method,and keeps considerable security.

关 键 词：CS-MRI 图像加密 Split-Bregman算法 数据降维 ARNOLD变换 

分 类 号：TP309.7[自动化与计算机技术—计算机系统结构]