CAJ | 학술논문

混沌对初值敏感的特性使得它适合于数据加密。以4阶CNN模型为基础，提出了一种新的超混沌细胞神经网络图像加密算法。算法分为置乱和扩散二个阶段，复合混沌映射用于生成置乱阶段控制参数，用以置乱图像行列之间的高度互相关像素。在扩散阶段，使用不同初始状态和参数的复合混沌映射生成高阶混沌细胞神经网络的初始条件，以生成扩散阶段的密钥流。算法的已知明文和选择明文攻击、密钥空间和直方图的仿真实验均取得了良好的结果。与其他相关算法相比，该算法具有密钥敏感性和抗攻击性强的优点，适用于图像加密。
혼돈대초치민감적특성사득타괄합우수거가밀。이4계CNN모형위기출，제출료일충신적초혼돈세포신경망락도상가밀산법。산법분위치란화확산이개계단，복합혼돈영사용우생성치란계단공제삼수，용이치란도상행렬지간적고도호상관상소。재확산계단，사용불동초시상태화삼수적복합혼돈영사생성고계혼돈세포신경망락적초시조건，이생성확산계단적밀약류。산법적이지명문화선택명문공격、밀약공간화직방도적방진실험균취득료량호적결과。여기타상관산법상비，해산법구유밀약민감성화항공격성강적우점，괄용우도상가밀。
Since chaos is sensitive for initial values, it is very suitable for data encryption. An image encryption algorithm based on hyper- chaotic control parameters and mixed scrambling diffusion structure of higher- order chaotic system was presented. The encryption algorithm included scrambling step and diffusion step. In the scrambling step, the composite chaotic map was used to generate the alignment phase control parameters and scramble for the high- level image cross- correlation between the adjacent pixels. In the diffusion step, the composite chaotic map with the different initial states and parameters was used to generate the initial conditions for hyper- chaotic cellular neural networks in order to generate the key stream. This method was evaluated by known plaintext attack and chosen plaintext attack, key space, image histogram, and simulations show good results. Compared with several other related algorithms, it has better anti- aggressive and key sensitivity is high. It can be applied to the image encryption.