CAJ | 학술논문

为了提高图像加密算法的安全性，利用细胞神经网络的混沌特性，结合分组加密结构，提出一种新的图像加密算法。首先对明文图像做2×2分块处理，然后转化为四组等长的数据流输入S盒进行置乱、替代操作；最后由S盒输出的四组数据流经分组加密网络加密后重构为密文图像。其中分组加密网络的密钥流是由三维细胞神经网络基于明文信息产生的。实验结果表明，细胞神经网络与分组加密网络能够有机融合形成安全性较高的图像加密算法。
위료제고도상가밀산법적안전성，이용세포신경망락적혼돈특성，결합분조가밀결구，제출일충신적도상가밀산법。수선대명문도상주2×2분괴처리，연후전화위사조등장적수거류수입S합진행치란、체대조작；최후유S합수출적사조수거류경분조가밀망락가밀후중구위밀문도상。기중분조가밀망락적밀약류시유삼유세포신경망락기우명문신식산생적。실험결과표명，세포신경망락여분조가밀망락능구유궤융합형성안전성교고적도상가밀산법。
We propose a new image encryption algorithm by using the chaotic characteristic of cellular neural networks and in combination with block cipher structure for improving the security of image encryption algorithm .First, the plaintext image is divided into 2 ×2 size blocks, each block is then transformed into 4 groups of data stream with equal length and sent to S-box for scrambling and substitution operation.Finally, four groups of data stream outputted from S-box are encrypted with block cipher network and reconstructed to ciphertext image.Moreover, the key stream in block cipher network is generated by 3D-cellular neural network based on plaintext information . Experimental results show that the cellular neural network and block cipher network can combine in an organic way to form image encryption algorithm with higher security .