忆阻耦合的三异质神经元网络及其图像加密应用
Memristor-coupled Neural Network with Triple Heterogeneous Neurons and Its Application in Image Encryption
王鹤鹏 1李志军1
作者信息
- 1. 湘潭大学 自动化与电子信息学院,湖南 湘潭 411105
- 折叠
摘要
探讨一个由三种异质神经元组成的环状神经网络的动力学行为及其在图像加密领域的应用.通过一个忆阻器模拟环形神经网络的两个神经元的耦合通道,探索神经网络在不同忆阻耦合强度下的非线性动力学特性.研究发现,在特定的耦合强度下,该神经网络展现出周期性和混沌放电行为的共存.基于其复杂且丰富的动力学特点,开发一种新型的图像加密算法.该算法结合DNA编码和RSA加密技术.选择DNA编码算法是因为其能够有效增强加密信息的随机性和不可预测性,而采用RSA加密算法是因为其基于大质数分解的难度,为加密过程提供了额外的安全性.该算法融合两种技术的优点,可显著提升加密过程的安全性和随机性,展示了高维神经网络在提高图像加密技术安全性方面的巨大潜力.
Abstract
This study explores the dynamical behavior of a ring-shaped neural network composed of three types of heterogeneous neurons and its application in the field of image encryption.By simulating the coupling channel between two neurons in the network using a memristor,the research explores the network's nonlinear dynamic characteristics under various memristor coupling strengths.It was found that at certain coupling strengths,the network exhibits a coexistence of periodic and chaotic discharge behaviors,revealing its complex and rich dynamic features.A novel image encryption algorithm is developed based on these findings.This algorithm combines DNA encoding and RSA encryption techniques.DNA encoding is chosen because it effectively enhances the randomness and unpredictability of the encrypted information,while the RSA encryption technique is used due to its reliance on the difficulty of large prime number factorization,providing additional security for the encryption process.This algorithm integrates the advantages of two techniques,significantly enhancing the security and randomness of the encryption process,while also demonstrating the tremendous potential of high-dimensional neural networks in improving the security of image encryption technology.
关键词
异质神经元/忆阻器/环形耦合/DNA编码/RSA加密Key words
heterogeneous neurons/memristors/ring coupling/DNA encoding/RSA encryption引用本文复制引用
基金项目
国家自然科学基金项目(62171401)
国家自然科学基金项目(62071411)
出版年
2024
NETL
NSTL
万方数据