摘要
利用离散Hopfield神经网络生成的序列密码,对图像信号进行加密和解密的实验研究表明,该网络可 模拟复杂非线性函数,使输出序列的周期性、0与1平衡性和序列相关性等均满足序列密码设计要求,而且并行运算 速度快,有利于数字电路实现.
This paper investigates a new image encryption method based on the stream cipher generated by a discrete Hopfield neural network.The experimental results show that the proposed discrete Hopfield neural network can be used to simulate complex nonlinear functions,and the period,0 and 1 equalization and sequence correlation properties of the output sequence can meet the requirements for the stream cipher design. Furthermore, the cipher can be easily implemented by digital circuits with parallel calculations to improve the processing speed.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2004年第4期677-679,共3页
Acta Electronica Sinica
基金
国家自然科学基金项目(No.60072012)
黑龙江省教育厅科技研究项目(No.10511015)
关键词
图像加密
移位寄存器
神经网络
序列相关性
image encryption
shift-register
neural network
sequence correlation
