基于卡方统计量的多差分攻击方法

为了精确地估计分组密码算法抵抗差分攻击的能力，在已知多个具有高概率差分特征的条件下，提出了基于卡方统计量的多差分攻击方法.分析了正确密钥和错误密钥对应的统计量的分布规律，给出了多差分攻击方法的成功率、数据复杂度和计算复杂度的关系.在分组密码算法的差分特征概率未知的条件下，该方法仍然是适用的.

Multiple Differential Cryptanalysis Using Chi-Square Statistics

In order to evaluate the capability that block ciphers resist differential attack accurately,a multiple differential cryptanalysis method is proposed in which chi-square statistics is constructed based on multiple differential characteristics with high probablities.We analyze the probability distribution of statistics corresponding correct key and incorrect key,and give the relation of data complexity,computational complexity and success probability.We point that the multiple differential cryptanalysis method can be applied to the the instance when the probabilities of differential characteristics are unknown.