中国组织工程研究

中国组织工程研究 ›› 2010, Vol. 14 ›› Issue (52): 9777-9780.doi: 10.3969/j.issn.1673-8225.2010. 52.022

• 数字化骨科 digital orthopedics • 上一篇    下一篇

Matlab程序实现心脏形态学形状主成分的聚类分析

侯园园，周  萍，张  韫   

  1. 首都医科大学生物医学工程学院，北京市  100069
  • 出版日期:2010-12-24 发布日期:2010-12-24
  • 通讯作者: 张韫，副教授，首都医科大学生物医学工程学院，北京市 100069 eduhelp@163.com
  • 作者简介:侯园园★，女，1985年，河南省安阳市人，汉族，首都医科大学生物医学工程学院在读硕士，主要从事医学图像处理研究。 hyy200333@163.com
  • 基金资助:

    首都医科大学基础临床基金（10JL51）。

Principal component of morphological traits of heart using Matlab: A cluster analysis

Hou Yuan-yuan, Zhou Ping, Zhang Yun   

  1. School of Biomedical Engineering, Capital Medical University, Beijing  100069, China
  • Online:2010-12-24 Published:2010-12-24
  • Contact: Zhang Yun, Associate professor, School of Biomedical Engineering, Capital Medical University, Beijing 100069, China eduhelp@163.com
  • About author:Hou Yuan-yuan★, Studying for master’s degree, School of Biomedical Engineering, Capital Medical University, Beijing 100069, China hyy200333@163.com
  • Supported by:

    the Basic Clinical Foundation of Capital Medical University, No. 10JL51e

PDF

462

摘要：

背景：快速的将心脏按其特征进行聚类可为后续统计分析和研究带来很大的便利。系统聚类法是将样品或变量按照其性质上的亲疏相似程度进行分类的一种多元统计方法。
目的：提出用主成分-聚类分析的方法来描述心脏形态学形状并进行分类，对中国健康成年人的心脏X射线测量的各项指标进行综合评价。
方法：搜集了36例健康成年人的胸片，并用MxLiteView软件手动测量了每幅胸片中代表心脏形态学形状常用的10个指标，用Matlab软件对测量指标进行主成分分析，然后对提取出的主成分进行聚类。
结果与结论：主成分分析后提取出3个主成分变量，将36例样本用提取的主成分进行聚类，可将样本分为3类，分别代表了心脏的3类不同的心型。用该方法对心脏形态学形状进行快速分类，对心脏的统计和分类研究提供了一定的参考价值。

关键词: 主成分分析, 聚类分析, 胸片, 心脏形态学, Matlab

Abstract:

BACKGROUND: Cluster of heart according to the characteristics can benefit subsequent statistical analysis and study. Cluster is a multielement method to classify sample or variable according to qualitative similarity.
OBJECTIVE: To propose the component-cluster method to describe and classify the shape of cardiac morphology and to comprehensively evaluate the heart indicators in the chest X-ray of Chinese healthy adults.
METHODS: We collected 36 chest X-rays of healthy adults, and manually measured 10 indicators, which represented the heart of morphological shapes using the software MxLiteView, We complied principal component analysis and cluster analysis using Matlab.
RESULTS AND CONCLUSION: Three principal components variables were extracted by principal component analysis. After cluster analysis using the extracted principal components, the 36 samples were divided into three categories, which represented three different types of heart shape. This method is used to classify the shape of heart morphology rapidly and provides some reference value for statistics and classification study of the heart.

中图分类号: