双随机相位加密系统的无约束最优化攻击

Unconstrained optimization attack on double random phase cryptosystem

提出一种针对双随机光学相位加密系统的无约束最优化攻击算法.在已知明文条件下,首次将双随机相位加密系统的攻击问题转化为一个单目标无约束最优化模型.基于该模型,在相应的攻击算法设计中,采用拟牛顿矩阵代替Hessian矩阵以准确获取系统的密钥,避免传统牛顿法需要计算Hessian矩阵的逆等严重缺陷.同时,因有效利用拟牛顿矩阵的正定、对称、可迭代求逆的特点,新的攻击算法具有恢复效果好、收敛速度快、初值依赖弱、鲁棒性较强等优势.此外,本算法所需约束条件较少,可方便地移植到其他光学加密系统的攻击中.

An unconstrained optimization method is proposed to attack the double phase encryption system. Under the condition of knowing the plaintext, the new attack method builds an unconstrained optimization model and gets the accurate phase key via this model. Using the acquired phase key, the attacker decrypts the followed cipher. The new attack method transforms the problem of attacking the double phase encryption system into an unconstrained optimization model. The new attack method replaces Hessian matrix by quasi-Newton matrix to avoid computation of the reverse of Hessian matrix. The new attack method has fast convergence speed and strong robustness, and it is not too sensitive to the original values of the variables. This attack method can be applied to other encryption systems.