CT图像重建的迭代算法

Iterative Algorithms in Computed Tomographic Reconstruction

随着计算机运算速度的飞速发展 ,以前只是被看成一种研究方法的迭代算法在图像重建中越来越引起人们的重视 ,并且已经在科研、工业和医学领域中得到了实际应用。本论文从CT图像重建原理入手 ,根据迭代重建的物理意义 ,从投影模型出发 ,得到投影数据 ,然后利用迭代算法MART及SIRT重建出断层图像 ,再与模型相比较 ,从而确定各算法的优劣 。

As the calculation speed of the CPU is fast developed, more consideration is paid to the iterative algorithms in tomographic reconstruction, which were only experimented as a research method before. Further more, it has been applied to practical use in such fields as science research, industry and medicine. The paper, based on the principle of image reconstruction and the physical meaning of iterative algorithms in tomographic reconstruction, starts from projection models, summarizes the optimization criteria of iterative reconstruction algorithms, MART and SIRT, makes comparisons between them, and then gives the prosperity how to select image algorithms in tomographic reconstruction. In order to approve the summarization, mathematical phantom is used to test the listed algorithms. During the test the projection y is obtained by processing the projections which are obtained from system function Radon transform. In addition, the projection matrix A is obtained by personal function, according to the structure of y.