基于Sugeno型模糊神经网络的外周血白细胞自动分类研究

Peripheral White Blood Cell Classification Based on Sugeno-model Fuzzy Neural Network

血白细胞分类计数嗜碱性粒细胞、嗜酸性粒细胞、中性粒细胞、淋巴细胞、及单核细胞是临床诊断的重要依据。作者在细胞图像分割、白细胞特征提取、白细胞特征选择后,提出了用Sugeno型模糊神经网络的方法实现外周血白细胞的自动形态学分类。用构建的网络分别进行白细胞的3分类和5分类试验,结果表明,用模糊神经网络分类器对血白细胞形态学分类可行。

Identifieation and classification of major white blood cell types named monoeytes, lymphoeytes, neutrophils, eosnophils, and basophils is important for clinical diagnosis. After cell segmentation, leukocytes＇ feature acquirement and selection, a classification method based on Sugeno- model fuzzy neural network was presented. The 3 - types classification and 5- types classification experiment results show that the fuzzy neural, network algorithm is feasible for white blood eell elassifieation.