摘要
本文报道了用于人体血液中六类白细胞自动分类的计算机分析系统，并提出了一套完整的识别算法．包括血液中细胞数字图象的采集、预处理、白细胞自动检出、区域分割、特征抽取和选择及分类判决．经大量作品实际测试表明：单个细胞检出成功率达９４％，六类识别率达８５％以上，三类达９０％以上．
This paper describes system for automatic classification of leukocytes and a series of algorithms, including the cell image segmentation, feature extraction and selection, and classifier's design. The classification accuracy is up to 85 percent for six categories and 90 percent for three categoried.The time for single leukocyte is less than two seconds.
出处
《计算机学报》
EI
CSCD
北大核心
1994年第2期130-136,共7页
Chinese Journal of Computers
关键词
白细胞分类
模式识别
图象处理
人
Leukocyte classification,pattern recognition,imageprocessing,mathematical morphology, texture analysis.
