血液细胞免疫特征模式识别技术研究

研究了药理实验中血液细胞免疫特征参数的模式识别技术 ,实现了药理实验的自动化 ,这对于药物的研究有着积极而深远的意义。介绍了血液细胞自动识别系统的组成及其处理流程 ,完成了复杂背景图像的分割。采用子图像细胞窗的方法检出单个细胞的子图像 ,缩小了图像处理区域 ,提高识别速度及准确率。分类统计血液细胞 ,实现了药理实验免疫特征参数的自动获取 ,解决了图像分割过程中被分成几个分离区域的单个细胞检出问题。测试结果表明 :药理参数的偏差不超过± 5 %。

Research on Pattern Recognizing Technique for Immunity of Blood Cells

This paper gives a new technique of automatically recognizing for immunity in pharmacological experiment. The realization of automatization of pharmacological experiment is very important and widely needed in pharmacology research and medical diagnosis. At the first, the paper introduces the system component and procession. The microscope image with complicated background is segmented. To improve the recognition speed and accuracy of the system, a way of dispatching image and multiprocessing for isolated cell is presented. The parameters of blood cells are obtained automaticly. The testing shows that the deviation of obtained pharmacological parameter doesn't exceed.