摘要
本文基于六个月孕妇MRI图像集并结合中国参考人数据,通过Matlab对图像进行预处理操作,使用手动分割和自动分割相结合的方法建立三维初始器官模型,再对初始器官进行优化并且生成NURBS曲面,通过逆向工程三维实体建模技术并依据解剖学分布数据对其进行装配构建起中国孕妇的曲面体模BREP-P6。该模型包含30多个器官,胎儿主要的构建部分有大脑、骨骼及软组织部分。分别计算了在五种不同标准入射条件下体外中子源(10−9 MeV~20 MeV)所致孕妇曲面模型中胎儿剂量转换系数,并将其结果跟RPI-P6模型进行对比。在五种入射条件下,随着中子源能量的增加,吸收剂量转换系数在两种模型的增长趋势基本一致,在10−9 MeV~1 MeV能量区域内,系数增长缓慢且误差较小,在1 MeV~20 MeV能量区域内,该系数增长较快。吸收剂量转换系数随着光子能量的增加,呈现正增长趋势。
In this paper, based on a six-month MRI image set of pregnant women and combined with Chinese reference human data, we used Matlab to pre-process the images, combined manual segmentation and automatic segmentation to build a 3D initial organ model, then optimized the initial organs and generated NURBS surfaces, and assembled them by reverse engineering 3D solid modeling tech-niques and based on anatomical distribution data to build surface model BREP-P6. The model con-tains more than 30 organs, and the main parts of the fetus are brain, bone and soft tissue. The fetal dose conversion coefficients in the surface model of pregnant women due to in vitro neutron sources (10−9 MeV~20 MeV) were calculated under five different standard incidence conditions, and the re-sults were compared with the RPI-P6 model. Under the five incident conditions, the growth trends of the absorbed dose conversion coefficients with increasing neutron source energy are basically the same in both models, with slow growth and small errors in the 10−9 MeV~1 MeV energy region and faster growth in the 1 MeV~20 MeV energy region.
出处
《应用数学进展》
2023年第5期2553-2560,共8页
Advances in Applied Mathematics
