基于混沌Hopfield神经网络视觉意义的图像加密算法

Visually meaningful image encryption algorithm based on chaotic Hopfield neural network

为解决传统加密方案得到的类噪声图像在传输过程中常常会因其视觉效果而受到攻击的问题,提出一种基于混沌Hopfield神经网络的具有视觉意义的双图像加密算法。对混沌Hopfield神经网络迭代生成随机数矩阵,与两幅压缩后的明文图像组合后进行离散余弦变换。通过生命游戏算法生成置乱矩阵来进行置乱。将置乱后的图像分为三部分,通过逆离散余弦变换分别嵌入到彩色载体图像的R、G、B分量中。实验结果表明,该算法密钥空间足够大,传输效率高,密钥敏感性强,能够抵御各种攻击,具有较好的安全性。

To solve the problem that noisy-like images obtained by traditional image encryption algorithms are usually attacked during transmission because of their visual effects,a visually meaningful double image encryption algorithm based on chaotic Hopfield neural network was proposed.The chaotic Hopfield neural network was iterated to generate a random number matrix.The matrix was combined with two compressed plain images and DCT transform was performed.The game of life algorithm was deployed to generate a matrix for permutation.The image was divided into three parts,and embedded into the R、G、B components of a color image by inverse DCT transform,separately.The results show that the key space of the proposed algorithm is big enough,and it has a great transmission efficiency,and is very sensitive to secret keys,as well as a great resistance to various attacks,thus it has a better security.