中文摘要：

针对超声CT重建问题,提出了一种基于高斯分布的最大似然期望最大化方法,并采用非最小最优化的方法来设置迭代的初始值。由于高斯分布更加符合测得时间信息的分布规律,所以基于高斯分布的新算法比传统基于泊松分布的最大似然期望最大化算法更加精确。所采用的非最小最优化方法能够减少迭代次数,有利于提高重建质量和计算效率。为了检验新算法,对三维温度场进行了仿真重建实验,结果表明,改进后的最大似然期望最大化算法具有更小的平均误差,能够得到更精确的重建图像。

英文摘要：

Aiming at the ultrasonic CT reconstruction problem, a maximum likelihood expectation maximization algorithm based on Gaussian model is proposed. And a non-minimization optimization method is presented to select proper initial values. Gaussian model is more suitable for the measured time information than Poisson model. Therefore the new maximum likelihood expectation maximization is superior to the original one. The non- minimization optimization method proposed in this paper can reduce the iteration times, and improve both reconstruction quality and computing efficiency. In order to validate the correctness of the proposed method, a 3D temperature reconstruction simulation is carried out. The results show that the improved maximum likelihood expectation maximization algorithm lowers the average error, and reconstruction images with better accuracy can be obtained.