摘要
详细介绍了一个基于极大似然法的骨髓细胞多光谱显微图像分类器的设计,给出了系统体系结构,细胞特征选择方法,分类器的训练及使用方法。实验结果表明,对训练细胞进行分类,绝大多数类别达到了90%以上的分类精度,同时分类速度也令人满意。
In this article,we particularly discusse a marrow cell high spectrum microscopic image classifier based on maximum likelihood method,giving the structure of the system,the feature selection method,the training and the using ways.The experimental result shows that the training cells were classified through the maximum likelihood classifier,most of the sorts have attained90%correct ratio with the satisfactory speed.
出处
《计算机工程与应用》
CSCD
北大核心
2005年第2期97-99,168,共4页
Computer Engineering and Applications
基金
"十五"科技攻关项目:光谱成像仪器的研制与开发
子课题名:光谱成像分析仪(编号:2001BA210A02)
关键词
多光谱
模式识别
分类器
极大似然估计
high spectrum,pattern recognition,classifier,maximum likelihood estimated value
