文章摘要

文章信息/Info

Title:: CT image reconstruction algorithm of region interest based on global variance

作者:: 陈飞1△; 冯晓刚2; 曲留更3; 1.周口市专科病医院放射科，周口 466000；2.河南省胸科医院放射科，周口 45000；3.周口市中心医院放射科，周口 46600

Author(s):: CHEN Fei1; FENG Xiaogang2; QU Liugeng3; 1.Department of Radiology,Zhoukou Provincial Hospital ,Zhoukou 466000,China;2.Department of Radiology,Henan Chest Hospital ,Zhoukou 466000;3.Department of Radiology,Zhoukou Central Hospital, Zhoukou 466000

Keywords:: Global variance; Region of interest; ART simulation; Iteration function; Sectional CT image; Reconstruction

摘要:: 感兴趣区域断层CT 图像重建在医疗诊断、工业无损检测领域广泛应用，重建精度直接影响医疗诊断结果与工业检测结果，为此提出基于全局方差的感兴趣区域断层CT图像重建算法。利用全局方差的断层CT图像增强算法步骤和流程，增强全局方差与全局均值的图像效果，同时获取区域内像素灰度级方差。在此基础上，导入ART模拟与投影，得到Radon逆变换ART迭代函数，重建增强效果后的感兴趣区域断层CT图像。研究结果表明，基于全局方差的重建算法可有效重建感兴趣区域图像，重建图像的归一化均方误差判距与归一化平均绝对距离判距均低于0.2，是一种可靠的感兴趣区域断层CT图像重建算法。

Abstract:: Region of interest（ROI） CT image reconstruction is widely used in medical diagnosis and industrial nondestructive testing. The reconstruction accuracy directly affects the results of medical diagnosis and industrial testing. To solve this problem, a region of interest tomography image reconstruction algorithm based on global variance was proposed. The global variance step and the flow chart of CT image enhancement algorithm were used to enhance the image effect of global variance and global mean, at the same time, the gray level variance of pixels in the region was obtained. On this basis, ART simulation and projection were imported to obtain the Radon inverse transform ART iteration function and reconstruct the enhanced ROI tomographic images. The results show that the reconstruction algorithm based on global variance can effectively reconstruct the ROI image, and the normalized mean square error and normalized mean absolute distance of the reconstructed image are less than 0.2, which is a reliable ROI tomography image reconstruction algorithm.