基于图像质量优化的锥束CT几何校正法  

A Geometric Calibration Method for Cone-beam CT Based On Image Quality Optimization

作　　者：杨帅[1] 齐宏亮[1] 吴书裕[1] 徐圆[1] 周凌宏[1] 

机构地区：[1]南方医科大学生物医学工程医疗仪器研究所,广州510515

出　　处：《核电子学与探测技术》2016年第7期697-701,共5页Nuclear Electronics & Detection Technology

基　　金：国家科技支撑计划项目(2015BAI01B10)资助

摘　　要：为了解决CBCT图像中因机械误差所导致的几何伪影的问题,提出一种基于图像质量优化的锥束CT几何校正方法。本方法通过引入CBCT图像几何伪影的评价指数,将评价指数作为衡量几何伪影严重程度的目标函数,通过迭代修正标定模体上标记点的坐标,进而达到校正几何参数和消除CBCT图像几何伪影的目的。仿真实验结果显示:CBCT系统的几何参数误差由19.07%降到2.21%,图像中的几何伪影明显减少;在真实实验数据中,该方法有效消除CBCT图像的几何伪影。本方法降低了标定模体加工误差对校正精度的影响,同时大幅度降标定模体的制作成本。In order to reduce the geometry artifacts caused by mechanical error in CBCT image,a geometric calibration method for cone-beam CT based on image quality optimization is proposed.This method introduced evaluation index which could evaluate the severity of geometric artifacts in CBCT image.Then setting the evaluation index as objective function which estimates the severity of geometric artifacts to amend the coordinates of marker points in calibration phantom by iterations,and further achieving the corrected geometry and amended CBCT image artifacts.The simulation results show that the streak artifacts are reduced and the error of geometry parameters decreases from 19.07% to 2.21%.In real experimental data,this method can effectively reduce the CBCT image geometry artifacts and the influence of the calibration phantom processing error for calibration accuracy,the expense of making accurate geometry phantom decreases.

关 键 词：锥束CT 几何校正 图像重建 标定模体 

分 类 号：TN911.73[电子电信—通信与信息系统]