[1]吴茜,皮一飞,周解平.CT/MRI混合配准方法及其在放疗计划系统中的应用[J].中国医学物理学杂志,2020,37(9):1148-1154.[doi:10.3969/j.issn.1005-202X.2020.09.013]
 WU Qian,PI Yifei,ZHOU Jieping.CT/MRI hybrid registration and its application in treatment planning system[J].Chinese Journal of Medical Physics,2020,37(9):1148-1154.[doi:10.3969/j.issn.1005-202X.2020.09.013]
点击复制

CT/MRI混合配准方法及其在放疗计划系统中的应用()
分享到:

《中国医学物理学杂志》[ISSN:1005-202X/CN:44-1351/R]

卷:
37
期数:
2020年第9期
页码:
1148-1154
栏目:
医学人工智能
出版日期:
2020-09-25

文章信息/Info

Title:
CT/MRI hybrid registration and its application in treatment planning system
文章编号:
1005-202X(2020)09-1148-07
作者:
吴茜1皮一飞2周解平3
1. 安徽医科大学人文医学学院,安徽合肥230032;2. 中国科学技术大学物理学院,安徽合肥230026;3. 中国科学技术大学附属 医院放疗科,安徽合肥230001
Author(s):
WU Qian1 PI Yifei2 ZHOU Jieping3
1. School of Humanistic Medicine, Anhui Medical University, Hefei 230032, China 2. Department of Physics, University of Science and Technology of China, Hefei 230026, China 3. Department of Radiation Oncology, Affiliated Hospital of University of Science and Technology of China, Hefei 230001, China
关键词:
多模态图像配准混合配准互信息B样条预处理
Keywords:
multi-modality image registration hybrid registration mutual information B-spline preprocessing
分类号:
R318
DOI:
10.3969/j.issn.1005-202X.2020.09.013
文献标志码:
A
摘要:
旨在研究放疗中图像配准方法,特别是针对放疗中常用的CT、MRI,提出基于混合框架的配准方法,该方法主要包 括两个方面:(1)采用掩膜(Mask)提取感兴趣区域、形态学运算等图像处理方法以及CPU多线程并行技术,大幅度提高配 准速度;(2)采用由全局到局部的混合配准策略,首先利用基于仿射变换的刚性配准整体配准,以防止图像间偏差过大,在 此基础上针对感兴趣区域采用B样条弹性配准,调整局部形变。通过实验表明,采用预处理及加速策略的刚性配准,在保 持其精度的情况下,提速比可达10倍,测试结果已达到临床需求;此外,采用基于GPU加速的混合配准策略,其配准速度 提至约4 min。
Abstract:
The methods for image registration in radiotherapy are investigated in the study. Aiming at CT and MRI commonly used in radiotherapy, a registration method based on hybrid framework is proposed. In the proposed method, image processing methods such as mask extraction of regions of interest and morphological operations as well as CPU multithreading parallel technology are used to greatly improve the registration speed, and a hybrid strategy of global and local registrations is adopted. Global rigid registration with an affine transformation is used to prevent the deviation between registered images, and then B-spline elastic registration is applied on regions of interest for adjusting local deformations. The experiments show that the preprocessing and acceleration strategy for rigid registration can increase the speed ratio by up to 10 times while maintaining its accuracy, and the test results reach the clinical requirements. In addition, the CT/MRI hybrid registration method based on GPU acceleration can achieve an average registration speed of 4 minutes.

备注/Memo

备注/Memo:
【收稿日期】2020-03-11 【基金项目】安徽医科大学博士科研资助基金(XJ201546);安徽高校 自然科学研究项目(KJ2019A0240) 【作者简介】吴茜,副教授,博士,主要研究方向:医学影像配准、医学 影像三维重建,E-mail: ayd_wuqian@126.com;皮一飞,博 士,主要研究方向:治疗计划系统研发、辐射防护剂量学, E-mail: 404963895@qq.com;周解平,硕士,主要研究方 向:临床肿瘤放疗,E-mail: 502709805@qq.com
更新日期/Last Update: 2020-09-25