中国组织工程研究

中国组织工程研究 ›› 2010, Vol. 14 ›› Issue (9): 1535-1538.doi: 10.3969/j.issn.1673-8225.2010.09.004

• 数字化骨科 • 上一篇    下一篇

基于多重分形的医学图像分割方法

金春兰，黄  华，刘圹彬   

  1. 四川大学电气信息学院，四川省成都市  610065
  • 出版日期:2010-02-26 发布日期:2010-02-26
  • 通讯作者: 黄 华，博士(后)，教授，博士生导师，四川大学医学信息工程系，四川省成都市 610065 hhua@scuedu.cn
  • 作者简介:金春兰，女，1985年生，朝鲜族，吉林省延吉市人，四川大学在读硕士，主要从事医学图像处理与医学信号处理研究。 Xiaoai20008@163.com

Medical image segmentation based on multifractal theory

Jin Chun-lan, Huang Hua, Liu Kuang-bin   

  1. School of Electrical Engineering and Information, Sichuan University, Chengdu   610065, Sichuan Province, China
  • Online:2010-02-26 Published:2010-02-26
  • Contact: Huang Hua, Doctor, Professor, Doctoral supervisor, School of Electrical Engineering and Information, Sichuan University, Chengdu 610065, Sichuan Province, China hhua@scuedu.cn
  • About author:Jin Chun-lan, Studying for master’s degree, School of Electrical Engineering and Information, Sichuan University, Chengdu 610065, Sichuan Province, China Xiaoai20008@163.com

PDF

491

被引次数

22

摘要：

背景：由于人体解剖结构的复杂性、组织器官形状的不规则性及不同个体间的差异性，所以比较适合用多重分形来分析。

目的：采用多重分形理论对医学图像进行图像分割。

方法：采用基于容量测度的多重分形谱计算及基于概率测度的多重分形谱计算方法对图像进行分割。对于待处理图片分别进行传统的区域生长分割，max容量测度图像分割，sum容量测度图像分割，概率测度图像分割等4种分割，并加入噪声后再进行同样的分割处理作为比较。

结果与结论：采用的两种基于多重分形谱的计算法中，基于容量测量的多重分形谱计算方法的关键是定义合适的测度μα；基于概率测度的多重分形谱计算方法的关键是定义合适的归一化概率Pi，不同的测度(概率)和不同的阈值对结果的影像比较大。基于概率测度的方法对噪声比较敏感，但是在滤过噪声时对图像象素大小变化比较大、比较复杂的图像有较好的分割效果。实验表明基于多重分形谱的医学图像分割方法在选择合适的测度(概率)和阈值时是可行的，特别是在较为复杂的图像处理中对于纹理和边缘的区别上有较大的优势，在准确地分割的同时能保留更多的细节，具有重要的实际意义。同时，多重分形也可以作为一种图像的特征，为特征提取多提供一种有力的数据。

关键词: 多重分形, 图像分割, 医学图像, 纹理, 数字化影像技术

Abstract:

BACKGROUND: As the complexity of human anatomic structure, the abnormity of tissue shape and the difference among individuals, the structure of multifractal is adapted.
OBJECTIVE: To investigate medical image segmentation based on multifractal.
METHODS: Image segmentation was performed by algorithm based on capacity measurement and probability measure. The experimental images were segmented using traditional region growing, max capacity measurement, sum capacity measurement, and probability measure. Following adding noise, the images were identically segmented and compared.
RESULTS AND CONCLUSION: In the two algorithms based on multifractal, the key of the algorithm based on capacity measurement is that appropriate measure μα is defined, and the key of the algorithm based on probability measure is that appropriate normalized probability Pi is defined. The different measures (probability) and thresholds bring greater effect. The method based on probability measure is sensitive to noises, but after filtration noise, segmentation effect is greater for the images whose pixels vary comparatively great and very complicated. The results show that it is feasible that appropriate measure (probability) and threshold is chosen based on medical image segmentation. Especially greater advantage exists for the distinction of texture and edge in the complicated image processing, which can reserve details while precisely dividing. It has very significant practical significance. At the same time, multifractal can also be characteristics of images, which provide powerful data for feature extraction

中图分类号: