感兴趣区域CT图像重建方法及模拟实验

Region-of-interest Image Reconstruction Algorithms and Numerical Experiments

虽然CT技术已经发展得相当成熟,但在大物体成像、高分辨率成像和减少辐射剂量等方面现有的CT成像方法仍然存在较大的困难。事实上,很多工程应用中并不要求对完整的物体进行全局CT成像,只需要获得某些感兴趣区域的物体图像即可,特别是医疗临床诊断中只要能够实现对可疑病灶部位的成像即可。因此,本文研究了针对感兴趣区域的CT图像重建方法,以及X射线束的视野只覆盖感兴趣区域的扫描方法设计,感兴趣区域CT成像研究的关键是如何在投影数据截断情况下实现断层图像的精确、稳定重建;本文介绍了三种投影数据在不同截断方式下的图像重建方法,并设计了相应的数值模拟实验,给出了数值模拟实验结果。最后,对感兴趣区域CT成像技术在工程中应用潜力做了展望。

Though CT technique has been developed more than forty years, there are many difficulties in large object CT imaging, high-resolution imaging and how to reduce X-ray dose. In fact, in stead of full images for the whole object, the images of region-of-interest （ROI） are enough in many CT imaging applications, especially in medical clinical imaging according to the lesion organs. This paper focuses on ROI image CT reconstruction algorithms and X-ray field-of-view design. The key problem of ROI reconstruction is how to exactly and stably reconstruct ROI images from the truncated projection data. This paper introduces reconstruction algorithms for three different kinds of truncation cases. Numerical experiments were done to validate the above algorithm. Finally, we discuss the application prospects of the ROI CT imaging technique.