基于密文随机性度量值分布特征的分组密码算法识别方案

在研究现有加密算法识别方案局限性的基础上，提出了基于密文随机性度量值分布特征的分组密码算法识别方案。首先，基于码元频数检测、块内频数检测及游程检测对AES、Camellia、DES、3DES及SMS4密文的随机性度量值取值个数进行了统计分析，采用k-means算法对其进行了初始聚类划分。其次，针对相同聚类中的分组密码算法识别问题，基于降低特征向量间相似度的原则，求解了码元频数检测、块内频数检测及游程检测对应的密文随机性度量值特征向量维数。最后，对AES、Camellia、DES、3DES及SMS4算法的实验结果表明，提出方案在已知密文条件下，实现了对以上典型分组密码算法的识别，相关成果可为进一步探索基于密文随机性度量值分布特征的加密算法识别提供参考。

Block ciphers identification scheme based on the distribution character of randomness test values of ciphertext

By researching deficiency of current encryption algorithms identification schemes, a block ciphers identification scheme is proposed based on the distribution character of randomness test values for ciphertext. Firstly, the numbers of randomness test values for AES, Camellia, DES, 3DES, SMS4 are respectively calculated based on the frequency test, frequency test in block, run test and originally clustered by the k-means algorithm. Secondly, in order to identify the block ciphers in each clustering, the dimensions of eigenvectors to the frequency test, frequency test in block, run test are calculated on the principle of reducing the comparability between eigenvectors. Eventually, the experimental results of AES, Camellia, DES, 3DES, SMS4 demonstrate that the proposed scheme effectively identified the above representative block ciphers, and the correlative research can promote the further encryption algorithms identification research.