基于压缩感知和分形图的视觉有意义的图像加密算法  

A visually meaningful image encryption algorithm based on compressive sensing and fractal graph

作　　者：杨宇光 王子嘉 YANG Yuguang;WANG Ziia(Faculty of Information Technology,Beijing University of Technology,Beijing 100124,China)

机构地区：[1]北京工业大学信息学部,北京100124

出　　处：《安徽大学学报（自然科学版）》2023年第5期32-43,共12页Journal of Anhui University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(62071015)。

摘　　要：基于压缩感知和分形图,提出一种视觉有意义的图像加密算法,从视觉安全和数据安全两方面为图像提供保护.明文图像哈希值与外部密钥共同生成混沌系统初始值,能有效提高算法抵御已知明文和选择明文攻击的能力.使用分数阶混沌系统与克罗内克积生成压缩感知所需的测量矩阵,增强安全性的同时能大幅缩短所需混沌序列的长度、提高算法的加解密效率.使用离散忆阻混沌系统控制分形图的生成、图像的置乱及按位异或操作,对压缩感知测量结果进行二次加密,在解决测量结果像素值分布不均问题的同时提高算法的安全性.实验结果表明:该文算法具有良好的视觉安全性,能有效抵抗穷举攻击、差分攻击、统计攻击等常见攻击;相对于其他文献算法,该文算法的保真度更高.Based on compressive sensing and fractal graph,a visually meaningful image encryption algorithm was proposed,providing image protection from both visual security and data security aspects.The plain image Hash value,combined with external keys,jointly generated the initial values of chaotic systems which could effectively improve the algorithm's resistance against known plaintext attacks and chosen-plaintext attacks.The fractional-order chaotic system and K ronecker product were used to generate the measurement matrices required for compressive sensing,could enhance security while significantly reducing the length of the required chaotic sequences and improving the efficiency of encryption and decryption.A discrete memristive chaotic system was used to control fractal graph generation,image scrambling,and bitwise XOR operation,enabling secondary encryption of the compressive sensing measurement result,could improve the algorithm's security while addressing the issue of uneven pixel value distribution in the measurement result.Experimental results demonstrated that the proposed algorithm exhibited good visual security,effectively resisted common attacks such as exhaustive attacks,differential attacks,and statistical attacks.Compared with other algorithms in literature,the proposed algorithm achieved higher fidelity.

关 键 词：视觉有意义的图像加密算法 压缩感知 分形图 分数阶混沌系统 离散忆阻混沌系统 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]