一种新的构建序列影像与呼吸信号对应关系的方法  

A novel method for constructing corresponding relationship between sequential image and respiratory signal

作　　者：崔学理 刘仰川[1] 夏威[1] 高欣[1] 

机构地区：[1]中国科学院苏州生物医学工程技术研究所,江苏苏州215163 [2]中国科学院长春光学精密机械与物理研究所,吉林长春130000 [3]中国科学院大学,北京100049

出　　处：《江苏大学学报（自然科学版）》2015年第2期187-190,共4页Journal of Jiangsu University：Natural Science Edition

基　　金：国家自然科学基金资助项目(81371640);江苏省科技计划项目(BL2012049);苏州市科技计划项目(SH201210)

摘　　要：构建肺部呼吸运动模型时,为了快速准确地确定序列影像在呼吸周期中的位置,提出了一种距离对应关联法,通过标记物相同时间段内在4D CT中和呼吸信号中的位移量的比较,确定4D CT和呼吸信号的对应关系.采用该方法建立双目立体视觉系统获取的呼吸信号和4D CT的对应关系,并以构建的肺部呼吸运动线性模型的模型误差验证方法的有效性.试验结果表明,基于距离对应关联法构建的肺部线性模型的模型误差在2.5 mm以内,相对于梯度对应关联法,模型误差降低超过10%;本方法简单易行、精度高,适合构建序列影像与呼吸信号间的对应关系.To quickly and accurately assign the positions of sequential images in respiratory cycle for building lung respiratory motion model,a novel distance correlation method was proposed to determine the corresponding relationship between 4D CT and respiratory signal by comparing the distance of in vitro labeling in 4D CT datasets with respiratory signal over the same time period. The proposed method was used to construct the correlational relationships between respiratory signal obtained by binocular visual system and 4D CT. The model error of a linear lung motion model built by the proposed method was calculated to evaluate the effectiveness of the proposed method. The error of the linear lung motion model based on the distance correlation method was within 2. 5 mm. Compared with the motion model based on the gradient correlation method,the error was reduced more than 10%. Experimental results show that the proposed method is feasible and effective with high accuracy,and it is suitable for constructing the relationship between sequential images and respiratory signal.

关 键 词：运动模型 序列影像 标记物 呼吸信号 对应关系 

分 类 号：TP391.7[自动化与计算机技术—计算机应用技术]