English

应用科学学报 ›› 1994, Vol. 12 ›› Issue (2): 120-126.

• 论文 • 上一篇    下一篇

血液白细胞计算机分类中的特征提取研究

杨晓敏, 罗立民, 韦钰   

  1. 东南大学
  • 收稿日期:1992-02-23 修回日期:1993-05-17 出版日期:1994-06-30 发布日期:1994-06-30

STUDY OF LEUKOCYTE FEATURES IN PERIPHERAL BLOOD FOR AUTOMATIC CLASSIFICATION YANG XIAOMIN

YANG XIAOMIN, LUO LIMIN, WEI YU   

  1. Southeast University
  • Received:1992-02-23 Revised:1993-05-17 Online:1994-06-30 Published:1994-06-30

PDF

129

被引次数

33

摘要： 在所建的血液白细胞自动分类LEUK分析系统上,对每个细胞提取了不同性质的三组共35个特征,包括7个形状参数、12个光密度参数和16个纹理参数。对1200余六类白细胞样品进行了自动分类测试。实验结果表明:若仅用形态和光密度特征,总体识别率达80%左右,其主要误识率来自于颗粒细胞的分类。而该文提出的纹理特征弥补这一不足,使系统的总体识别率提高了十个百分点。该文重点介绍这三组特征,并就其在白细胞自动识别分类中的作用进行详细讨论。

关键词: 图象处理, 纹理分析, 白细胞分类, 特征提取

Abstract: For each cell,a total of 35 features in three kinds are extraeted from the automatic classification system of peripheral blood leukocytes.They are:7 morpho-logic features,12 optical densiey features and 16 texture features.The testing for 1200 leukocytes indicates that using these three kinds of features is better thanjust using the first 2 kinds,and the accuracy rate is approximately 90%. The result is discussed.

Key words: leukocyte classification, feature extraction, texture analysis, image processing