摘要
考虑到支持向量机(SVM)在训练样本有限的情况下处理高维数据上的优势,鉴于白细胞多光谱图像数据维数高的特点,为提高白细胞识别的速度和精度,采用支持向量机对白细胞的多值分类问题进行了研究,设计并实现了核函数为二值径向函数(RBF)的分类器,实验结果表明,该分类器有效地解决了白细胞的识别速度和精度问题,识别率达到了89.02%。
In view of support vector machine （SVC） is validated to resolve the recognition problem which swatch is small and the dimension of the feature space is high, because of the high-dimension data of multi-spectroscopy microscope white blood cell image, for improving the ration and speed of white blood cell recognition, the way of using SVC to realize white blood cell＇s auto recognition is introduced. A classifier is realized, it uses RBF as sorting function. The result shows the method is validity, the ration of recognition can reach 89.02%.
出处
《计算机工程与设计》
CSCD
北大核心
2008年第1期184-186,共3页
Computer Engineering and Design
基金
教育部开放式基金项目(K200611)
关键词
多光谱图像
特征选择
白细胞
支持向量机
模式识别
spectral image
feature selection
white blood cell
support vector machines
pattern recognition
