摘要
通过对目标连通区域标记及紧凑度、重心位置计算的方法来实现数字图像中特征目标自动识别的功能.并将其应用到医学图像中统计分析血液细胞的特征和物性确定细胞是否由病变.实验结果表明,该方法在血液分析中实用有效.
This paper presents how to automatically recognize components in a digital iconological image by labeling connected component and calculate its compactness and centriod . Then this method was applicated in iconological image processing for analyzing blood cell. The experiment results show that the method is effective for analyzing blood samples.
出处
《华东交通大学学报》
2005年第5期82-85,共4页
Journal of East China Jiaotong University
关键词
连通区域
区域标记
紧凑度
connected component
component labeling
compactness
