Abstract: By analyzing the current image encryption algorithm based on a class of artificial neural network in recent years, this paper proposes a new hyperchaotic image encryption algorithm based on 4D of Hopfield-type neural network and AES（Advanced Encryption Standard） encryption algorithm. Firstly, it defines four numbers and a average value of plaintext image pixel as key, hyperchaos sequence generated by Hopfield-type neural network as the goal keys of AES algorithm. Then, the plaintext image and hyperchaos sequence do exclusive or processing. Finally, the goal key is generated into the AES algorithm of triple encryption cipher. The simulation results show that the algorithm can effectively combine the advantages of two algorithms, achieving good encryption effect.

Key words: 4D of Hopfield-type neural network, hyperchaotic, Advanced Encryption Standard（AES） encryption algorithm

摘要： 针对近几年人工神经网络在图像加密领域的应用进行了研究，提出了一种新的四维Hopfield神经网络和AES（高级加密标准）加密算法相结合的超混沌图像加密算法。该方法首先定义了四个数和提取一个明文像素值的平均值作为密钥，通过Hopfield神经网络生成的超混沌序列作为AES加密算法的目标密钥；然后，将明文与目标密钥进行异或处理；最后，将目标密钥代入算法进行三重AES加密算法进行加密得到密文。通过仿真实验表明，该算法能够很好地结合两种算法的优点，达到非常好的加密效果。

关键词: 四维Hopfield神经网络, 超混沌, 高级加密标准（AES）加密算法