次梯度投影算子求解图像重建问题的并行策略

Parallel strategy of subgradient projection operator for image reconstruction

将次梯度投影迭代算法应用到数字图像重建问题。将图像重建问题转化为求一个加权最小二乘问题,导出次梯度投影算子在该问题下的具体迭代形式,并采用并行计算策略重建算法。通过三维数值实验对比次梯度投影迭代算法与常用的SART算法,验证算法的可行性和效率。

The subgradient projection iterative algorithm is applied to digital image reconstruction. The problem of image reconstruction is transformed into a weighted least-square problem,the specific iterative form of the subgradient projection operator is derived,and parallel computing strategy reconstruction algorithm is adopted. The iterative algorithm of subgradient projection and the common SART algorithm are compared by three-dimensional numerical experiments. The feasibility and efficiency of the algorithm are verified.