细胞自动计数中的一种图象标号算法

An Image Labeling Algorithm Used in the Auto-counting of Cells

在临床诊断和医学科研中,很大部分的工作与细胞的计数有关。传统的人工细胞计数方法由于其繁琐乏味和低效性而受到诸多限制,本文在高性能图象处理系统中,采用特定的图象处理方法——图象标号算法,完成一般病理切片的细胞计数,克服人工方法的上述不足。本文首先简单介绍了完成细胞图象分析的图象处理系统和图象分析过程,然后较详细地描述了细胞自动计数中所采用的图象标号算法,还用实际临床病理切片对该方法进行了实验,并用照片和图表给出实验结果。在本文的最后,对该标号算法的有效性进行了相关的讨论。

Almost all the cell microscopic imago analyses in medical clinics and studies are based on cell counting. However, this traditional tedious manual work usually takes a great deal of time and labor inefficiently and inaccurately. With the helps of a high performance image processing workstation and advanced processing technologies, it is possible for us to find an efficient automatic cell counting method based on image processing to solve all the problems stated above. Efficiency and accuracy of cell counting are chiefly considered.In this paper, a system performing cell image analyses is simply introduced at first, following the description of the image analysis procedures. Then the image labeling algorithm used in the auto-counting of cells is described in more details. Furthermore, this method is tested on a clinical pathological microscopic image and the experimental results are given with photos and tables. sIn the end of this paper, a discusion on this algorithm is taken and its results on an arbitrary case are given.