密集型细胞显微图像高精度快速计数方法  

Method of High Precision and Fast Counting for Intensive Cell Microscopic Image

作　　者：程揭章 嵇晓强[1] 李明光[1] 

机构地区：[1]长春理工大学生命科学技术学院,长春130022

出　　处：《长春理工大学学报（自然科学版）》2014年第2期71-75,共5页Journal of Changchun University of Science and Technology(Natural Science Edition)

基　　金：吉林省科技厅项目(20121006)

摘　　要：细胞显微图像计数在现代生物医学领域发挥着非常重要的作用。本文基于数字图像处理技术,研究一种密集型细胞显微图像高精度快速计数方法。针对密集型细胞重叠,粘连较多,细胞间距不明显等特点,提出了一种快速连通域标记算法及基于数学统计学的细胞计数方法。大量实验验证结果表明:该方法计数精度能达到95%以上,弱化了由图像分割不到位造成的统计误差,可应用于多个不同种类细胞的计数,通用性较强。并且由于结合了快速图像预处理及分割优化方法,极大缩短了算法运行时间,可满足一般工程应用中细胞实时计数的要求。Cell microscopic image count plays an important role in modern biomedicine. In this paper, an fast counting method with high accuracy of intensive cell microscopic image based on digital image processing techniques is re-searched. Aiming at the features that the intensive cell features are overlapped,adhesion and spacing are not obvious,a fast connected component labeling algorithm and a method of cell counting based on mathematical statistics are put for-ward in consider of the real-time performance. A large number of experimental results show that the method of count-ing accuracy can reach more than 95%. The statistical error caused by image segmentation which is not in place is weakened. This method can be applied in many different types of cell counting and have strong universal.As a combina-tion of fast image and segmentation method with optimization,the running time of the algorithm is shortened greatly by using this method,which can meet the engineering application requirements in real-time cell counting.

关 键 词：密集型细胞计数 图像分割 连通域标记 统计学 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]