Home > Browse Issues > Vol.39 No.11
Research on Automatic Analysis Algorithm of Mesenchymal Stem Cell Growth Confluence
Bai Hua1,2*, Zhang Fengfeng1,2, Zhang Cheng1,2, Zhao Junfa1,2, Yan Shulin3,4, Zhang Jianzhong3,4, Han Zhibo3,4
1College of Electronics and Information Engineering, Tianjin Polytechnic University, Tianjin 300387, China; 2Tianjin Key Laboratory of Optoelectronic Detection Technology and System, Tianjin 300387, China; 3National Engineering Research Center of Cell Products, Tianjin 300457, China; 4Tianjin Amcellgene Engineering Co., Ltd., Tianjin 300457, China
Abstract: The cell confluence is an important parameter in cell culture in vitro. Currently， the assessment of this parameter is usually carried out by human， and thus there are many defects such as inefficiency， poor precision and low reliability. Image processing technology is a powerful tool that has been widely used in biomedical field owing to its characteristics of high-speed， high-veracity and high-automation. Therefore， the use of this technique for cell confluence analysis can greatly improve the detection efficiency and objective accuracy. An image processing method for automatic analysis of adherent cell confluence is presented in this paper. Firstly， the cell microscopic images were preprocessed by the top-bottom hat of gray morphology and background subtraction. Secondly， an improved K-Means clustering algorithm was used to coarsely distinguish between foreground and background. Then， an algorithm combining area filtering and binary morphology was adopted to optimize segmentation effect， which could obtain a good segmentation between cellular area and background， and finally calculated the cell confluence. Experimental results showed that the algorithm had a high accuracy and could effectively replace the traditional artificial method to automatically detect the cell confluence.
CN
EN