计算机工程与应用 ›› 2015, Vol. 51 ›› Issue (12): 171-175.

• 图形图像处理 • 上一篇    下一篇

基于Kmean和ELM的乳腺肿块检测方法

王梦珍,刘  立,王  建   

  1. 天津大学 电子信息工程学院,天津 300072
  • 出版日期:2015-06-15 发布日期:2015-06-30

Detection of masses in mammograms using Kmeans and Extreme Learning Machine

WANG Mengzhen, LIU Li, WANG Jian   

  1. School of Electronic Information Engineering, Tianjin University, Tianjin 300072, China
  • Online:2015-06-15 Published:2015-06-30

摘要: 肿块是乳腺癌在X线图像上的一个主要表现。提出了一种肿块自动检测算法。该方法包括四个步骤:在图像预处理阶段,去除背景、标记、胸肌和噪声,图像分割和图像增强;利用Kmean方法找到感兴趣区域(ROI);提取能够表征肿块的特征;利用极限学习机(Extreme Learning Machine,ELM)分类器去除假阳性,将图像中的肿块和非肿块分离开来。通过对MIAS数据库中乳腺X线图像的测试实验,得到的检测肿块的准确率为93.5%。

关键词: 乳腺肿块检测, Kmean, 特征提取, 极限学习机(ELM)

Abstract: Mass detection on mammography is an effective method for breast cancer diagnoses. An automated mass detection method is proposed, including four steps: mammograms are preprocessed to remove background, tags, pectoral muscles; K-mean method is used to segment the Region Of Interest(ROI); features of mass such as shape and texture are extracted; a machine learning method is then applied to identify masses from ROIs using features. The mammograms of MIAS database are used for testing, and the accuracy of the masses detection is 93.5%.

Key words: detection of masses, Kmean, feature extraction, Extreme Learning Machine(ELM)