|
|||||
|
|
| 基于支持向量机的CD4细胞图像识别方法 | |
| 引用本文： | 张问银,金宁德,刘印锋.基于支持向量机的CD4细胞图像识别方法[J].计算机工程与科学,2009,31(7). |
| 作者姓名： | 张问银 金宁德 刘印锋 |
| 作者单位： | 1. 天津大学电气与自动化工程学院,天津,300072;临沂师范学院信息学院,山东,临沂,276005 2. 天津大学电气与自动化工程学院,天津,300072 3. 临沂师范学院信息学院,山东,临沂,276005 |
| 基金项目： | 博士后基金资助项目，国家自然科学基金资助项目 |
| 摘 要： | 本文基于支持向量机(SVM)对CD4细胞图像进行了分析识别。在样本数较少和缺乏先验知识的情况下,采用改进的平行栅格搜索算法和交叉验证算法训练并确定分类器参数,提高了分类器性能。实验结果表明,基于支持向量机的分类器的识别率为95%,优于Fisher分类器和BP网络。 |
| 关 键 词： | CD4细胞 显微图像识别 支持向量机 SVM |
SVM-Based Recognition of the CD4 Cell Images |
|
| ZHANG Wen-yin,JIN Ning-de,LIU Yin-feng.SVM-Based Recognition of the CD4 Cell Images[J].Computer Engineering & Science,2009,31(7). | |
| Authors: | ZHANG Wen-yin JIN Ning-de LIU Yin-feng |
| Affiliation: | 1.School of Electrical Engineering and Automation;Tianjin University;Tianjin 300072;2.School of Information;Linyi Normal University;Linyi 276005;China |
| Abstract: | A method to analyse and recognize the microscopic CD4 cell image based on support vector machines (SVM) is introduced with a small sample size and less prior knowledge in this paper.The method uses an improved parallel grid search algorithm to choose the parameters of the SVM classifier,and employs cross validation to test the performance of SVM classification. Using this new method,an accuracy of 95% has been obtained in the classification of the CD4 cell images,suggesting that the SVM approach is better t... |
| Keywords: | SVM |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ | |
