作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2008, Vol. 34 ›› Issue (2): 14-16. doi: 10.3969/j.issn.1000-3428.2008.02.005

• 博士论文 • 上一篇    下一篇

基于形态特征和SVM的血液细胞核自动分析

曾 明1,孟庆浩1,张建勋2,鲍菁丹1   

  1. (1. 天津大学电气与自动化工程学院,天津 300072;2. 南开大学机器人与信息自动化研究所,天津 300071)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2008-01-20 发布日期:2008-01-20

Automatic Analysis System of Blood Cell Nuclei Based on Morphological Features and Support Vector Machines

ZENG Ming1, MENG Qing-hao1, ZHANG Jian-xun2, BAO Jing-dan1   

  1. (1. School of Electrical Engineering and Automation, Tianjin University, Tianjin 300072;2. Institute of Robotics & Automatic Information System, Nankai University, Tianjin 300071)
  • Received:1900-01-01 Revised:1900-01-01 Online:2008-01-20 Published:2008-01-20

摘要: 以形态学分析和支持向量机为基础,构建了一套血细胞核显微图像自动分析与识别系统。在细胞核分割阶段,提出基于支持向量机的血液细胞核彩色图像分割算法。在特征提取环节中,除使用常规形态特征外,提出了一种新的能有效反映核分叶数差异的形态特征——腐蚀退化因子。采用“one-against-one”策略的多分类SVM方法对血细胞进行分类识别。实验测试表明,该系统具有较高的识别精度,平均识别率达94.13%。

关键词: 血液细胞核, 图像分割, 支持向量机, 腐蚀退化因子

Abstract: Based on morphological analysis and Support Vector Machines (SVM), a robust automatic analysis system of blood cell nuclei is developed. A novel algorithm for color image segmentation of blood cell nuclei based on the SVM is proposed. A new morphological feature named erosion degenerate factor is used to indicate the lobulated state of cell and achieves feature extraction by combining with traditional characteristics. One-against-one multi-class SVM is applied to classify the blood cell. Experimental results show that the proposed system yields better performance with the average recognition rate of 94.13%.

Key words: blood cell nuclei, image segmentation, Support Vector Machines(SVM), erosion degenerate factor

中图分类号: