基于特征提取量化分析的体外活细胞追踪算法研究

Algorithm for tracking living cells in vitro based on quantitative analysis of characteristic extraction

目的:提高显微镜下序列图像细胞追踪的效率及准确度。方法:提出双阈值形态学与拓扑约束图论法相结合的自动细胞追踪算法,用来分析体外活细胞定向迁移轨迹及参数,并从细胞数目及细胞特征两方面分析追踪算法的准确性。在特征分析方面,从运动速度、运动距离、趋化速度、趋化指数和方向持续性5个指标与手动采样数据进行对比。结果:该算法可以分别识别在毛细管针部灰度较高区域的细胞及其他区域灰度较低的细胞,细胞数目准确度平均达到91.8%,分析得到的5个特征指标与手动采样分析结果基本一致,误差不超过5%。结论:双阈值形态学与拓扑约束图论法相结合的自动细胞追踪算法可以有效提高细胞追踪的准确度。

Objective To improve the efficiency and accuracy of cell tracking algorithm for sequential images under a microscope.Methods An automatic cell tracking algorithm based on two-step threshold and morphology with topology-graph theoretical approach was proposed to analyze the trajectories and motility parameters of living cells in vitro.The accuracy of the cell tracking algorithm was analyzed in the aspects of cell number and cell characteristics.In characteristic analysis,5 characteristic parameters were compared between automatic cell tracking and manual cell tracking,including motility distance,motility speed,chemotaxis speed,chemotaxis index,and persistency.Results The proposed algorithm achieved the identification of cells in high gray-level regions in the capillary tube and those in other lower gray-level regions,and the accuracy in cell number was 91.8%.The 5 characteristic parameters of automatic cell tracking were consistent with the results of manual sampling,with an error not exceeding 5%.Conclusion The automatic cell tracking algorithm based on two-step threshold and morphology with topologygraph theoretical approach can effectively improve the accuracy of cell tracking.