首页 | 本学科首页   官方微博 | 高级检索  
     

一种利用互信息的多核并行医学图像配准算法
引用本文:马永军,吴腾华,孙嘉栋,田婕妤.一种利用互信息的多核并行医学图像配准算法[J].天津轻工业学院学报,2010(6):62-66.
作者姓名:马永军  吴腾华  孙嘉栋  田婕妤
作者单位:[1]天津科技大学计算机科学与信息工程学院,天津300222 [2]天津科技大学电子信息与自动化学院,天津300022
基金项目:天津市科技支撑计划重点资助项目(08ZCKFGX00600)
摘    要:分析了利用互信息作为相似性测度进行医学图像配准的算法,给出了具体计算流程.针对其中互信息计算量大、耗时长的缺点,提出了一种运行于单机多核平台的快速并行配准算法.利用OpenMP(open multi-processing)构建了一个图像匹配的多核并行计算平台,并对配准程序中的互信息计算进行并行处理,最后完成配准.通过对图像匹配算法效率进行的评估实验,验证了多核并行计算技术能够提高医学图像配准的运行效率.结果表明,该方法既保证了配准精度,又能够较好解决配准速度慢的问题.

关 键 词:医学图像  快速配准  并行  互信息

An Algorithm of Multi-Core Parallel Medical Image Registration Using Mutual Information
MA Yong-jun,WU Teng-hua,SUN Jia-dong,TIAN Jie-yu.An Algorithm of Multi-Core Parallel Medical Image Registration Using Mutual Information[J].Journal of Tianjin University of Light Industry,2010(6):62-66.
Authors:MA Yong-jun  WU Teng-hua  SUN Jia-dong  TIAN Jie-yu
Affiliation:1.College of Computer Science and Information Engineering,Tianjin University of Science & Technology,Tianjin 300222,China;2.College of Electronic Information and Automation,Tianjin University of Science & Technology,Tianjin 300222,China)
Abstract:Due to the inferiority of low speed and huge calculation in medical images registration based on mutual information,a fast registration algorithm was discussed.In the image registration process,the algorithm adopts two-step search strategy to target range.An OpenMP based parallel computing platform was built by using multi processor cores,where the parallel processing of the loop in the registration program is on the single-chip double-processor.Experimental results show that the method can solve the problem of low speed,which keep precision of registration.
Keywords:medical image  rapid registration  parallel  mutual information
本文献已被 维普 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号