粒子群算法多目标优化下的超混沌人脸图像加密

Multi-objective optimization of hyperchaotic face image encryption based on particle swarm optimization algorithm

中文关键词  混沌系统  粒子群算法  图像加密  智能优化  人脸隐私保护

英文关键词  chaotic system  particle swarm optimization algorithm  image encryption  intelligent optimization  face privacy protection

中文摘要

      本文将粒子群优化算法(PSO)与超混沌系统相结合, 提出一种基于多目标优化的人脸图像加密方案. 该方 案通过PSO算法协同优化多项加密评估指标, 包括相关关系、像素变化率(NPCR)、统一平均变化强度(UACI)和信息 熵. 首先, 初始化混沌系统的控制参数, 并采用SHA-256算法生成混沌系统的初始值, 迭代生成高敏感性的随机序 列; 其次, 利用随机序列执行像素置乱、扩散和行列置乱操作, 生成初始加密人脸图像; 然后, 将加密人脸图像视为 PSO算法的个体, 通过迭代更新个体的位置优化考虑多项指标的适应性函数; 最后, 确定混沌系统的最优参数, 并得 到最佳的加密人脸图像. 实验结果表明, 本文的方法在信息熵、像素相关系数、NPCR和UACI方面的表现都优于主 流方法, 这说明本文所提方法具有更高的安全性.

英文摘要

      In this paper, a particle swarm optimization algorithm (PSO) is combined with a hyperchaotic system to present a face image encryption scheme based on multi-objective optimization. The scheme co-optimises various encryption evaluation metrics, including pixel correlation, number of pixel change rate (NPCR), uniform average changing intensity (UACI) and information entropy through the PSO algorithm. Firstly, the control parameters of the chaotic system are initialized, and the initial values of the hyperchaotic system are generated using the SHA-256 algorithm to iteratively produce highly sensitive random sequences. Secondly, the random sequences are used to perform pixel permutation, diffusion, and row-column permutation operations, resulting in the initial encrypted face image. Then, the encrypted face image is treated as an individual of the PSO algorithm, and the fitness function considering multiple metrics is optimized by iteratively updating the individual’s position. Finally, the optimal parameters of the hyperchaotic system are determined, and the best encrypted face image is obtained. Experiments demonstrate that the proposed algorithm outperforms the mainstream methods in terms of information entropy, pixel correlation coefficient, NPCR, and UACI, which indicates that the proposed method has higher security.