扰动因素影响下对修正的OS-EM算法的收敛分析

Convergence analysis of the modified OS-EM algorithm for noisy data

在扰动因素影响下,通过引入松弛参数对Ordered-Subsets Expectation-Maximization(OSEM)算法进行修正.基于修正迭代的L^1收敛性,我们通过修改相应停止规则中的参数得到了一些主要的单调性结果,进而证明了迭代的L^2收敛性.

In the noisy data case, we modify the Ordered-Subset Expectation-Maximization（OS-EM） algorithm by introducing a relaxation parameter. Based on the convergence of modified iteration in L^1 norm, we obtained some main monotonic results by modifying the parameters of the corresponding stopping rule, and proved the convergence for this modified iteration in L^2 norm.