融合神经网络变换和通道置乱的彩色图像加密  

Color image encryption based on neural network transformation and channel scrambling

作　　者：栗风永 魏璐 曾祎姝 LI Feng-yong;WEI Lu;ZENG Yi-shu(College of Computer Science and Technology,Shanghai University of Electric Power,Shanghai 201306,China)

机构地区：[1]上海电力大学计算机科学与技术学院,上海201306

出　　处：《计算机工程与设计》2023年第7期2118-2124,共7页Computer Engineering and Design

基　　金：上海市自然科学基金项目(20ZR1421600);国家自然科学基金通用技术联合基金项目(U1736120)。

摘　　要：针对传统彩色图像加密过程中密钥空间小、彩色通道加密不彻底导致的密文图像抗差分攻击能力弱等问题,提出一种融合神经网络随机变换和通道置乱的彩色图像加密算法。将彩色图像R、G、B色彩通道进行分离,利用变换技术分别对3个色彩通道进行像素级置乱;利用并行混沌系统生成两组伪随机序列,引入人工神经网络对其进一步变换生成新序列。使用随机序列分别对彩色图像的通道内和通道间进行像素级扩散和置乱。实验结果表明,神经网络随机性可显著扩大密钥空间,提出算法在密文图像安全性和抗攻击能力上具有显著优势。Aiming at the problem of small key space and incomplete color channel encryption in existing color image encryption schemes,a color image encryption algorithm was proposed by combining neural network random transformation and channel scrambling.The color image was separated into R,G and B colorful channels,and traditional transformation techniques were employed to implement pixel-level scrambling.Subsequently,parallel chaotic system was defined to produce two sets of pseudo-random sequences.An artificial neural network was introduced to further transform the pseudo-random sequences to generate new encryption sequences.The new generated sequences were used to achieve pixel diffusion and color channel scrambling of color images.Experimental results show that the neural network can effectively expand the random key space,the proposed scheme obtains significant advantages in terms of encryption security and anti-attack capability.

关 键 词：彩色图像加密 色彩通道 像素置乱 混沌系统 神经网络 随机序列 像素扩散 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]