|本期目录/Table of Contents|

[1]万振环.一种适用于肝脏CT图像配准改进的尺度不变特征变换算法*[J].生物医学工程研究,2019,03:326-330.
 WAN Zhenhuan.An improved scale-invariant feature transform algorithm for liver CT image registration[J].Journal of Biomedical Engineering Research,2019,03:326-330.
点击复制

一种适用于肝脏CT图像配准改进的尺度不变特征变换算法*(PDF)

《生物医学工程研究》[ISSN:1006-6977/CN:61-1281/TN]

期数:
2019年03期
页码:
326-330
栏目:
出版日期:
2019-09-25

文章信息/Info

Title:
An improved scale-invariant feature transform algorithm for liver CT image registration
文章编号:
1672-6278 (2019)03-0326-05
作者:
万振环
厦门医学院, 厦门 361000
Author(s):
WAN Zhenhuan
Xiamen Medical College, Xiamen 361000,China
关键词:
肝脏CT图像尺度不变特征变换配准分块
Keywords:
Liver CT Images Scale-invariant feature transform Registration Regional block
分类号:
R318;TP391
DOI:
10.19529/j.cnki.1672-6278.2019.03.13
文献标识码:
A
摘要:
针对肝脏CT图像特点,在传统的尺度不变特征变换(scale-invariant feature transform ,SIFT)算法基础上,结合K-means聚类算法,提出了一种改进的特征点匹配算法。该算法通过聚类SIFT特征点坐标,将配准图像分为4个区域,特征点分块配准。与原算法相比,该算法增加了特征点匹配数量,有效隔离了特征点跨区域的错误匹配,时间复杂度也得到了一定的降低。该算法还减少了肝脏CT图像配准中错误匹配对配准结果的影响,提升了肝脏CT图像的配准精度。
Abstract:
According to the characteristics of liver CT images, an improved feature point matching algorithm was proposed based on the traditional scale-invariant feature transform(SIFT) algorithm and combining with K-means clustering algorithm. By clustering SIFT feature point coordinates, the registration image was divided into four regions and feature points were segmented for registration. Compared with the original algorithm, this algorithm increases the number of feature point matching, effectively isolates the cross-region mismatch of feature points, and reduces the time complexity. The algorithm reduces the influence of mismatched liver CT image registration results and improves the accuracy of liver CT image registration.

参考文献/References

备注/Memo

备注/Memo:
(收稿日期:2019-01-20)福建省教育厅中青年教师教育科研项目(JT180641);厦门医学院教改项目(XBJG2018003)。 Email:wzh@xmmc.edu.cn
更新日期/Last Update: 2019-10-24