甲襞毛细血管的自动跟踪研究

STUDY ON SELF-TRACKING OF NAILFOLD CAPILLARIES

本文分析了甲襞毛细血管图像的特性，据此提出了甲襞毛细血管的软件自动跟踪窗口技术：即在时序图像上以视频速度搜索定位甲襞毛细血管，然后实施采样存贮。在计算机及图像输入设备给定的情况下，采样及存贮毛细血管图像窗口（其尺寸恒定）所需时间恒定，因而为了以视频速度跟踪甲襞毛细血管就必须将搜寻毛细血管所花时间降至最小。采用了两条措施来降低搜寻毛细血管的时间。第一条措施是，建立了甲襞毛细血管自动跟踪的数学模型，进而建立了ｆ函数，这种模型易于检测及搜索。第二条措施是，设计了最佳搜索算法，保证以最短时间搜索到毛细血管。最终实现了甲襞毛细血管的自动跟踪。

Based on the analysis of characteristics of nailfold capillay images,a software-based self-tracking window technique for nailfold capillary is presented.This technique constituted searching for capillaries in temporal image series and sampling and storing of capillaries at video rate.Since th etime needed for sampling and storing of capillaries.Since the sampling and storing of capillaries is constant under given conditions of computer and image input device,it is of crucial importance to reduce the time needed for searching capillaries in order to track the capillaries at video rate.Two means were used to reduce capillary searching time.A mathematical model convenient for self-tracking of nailfold capillaries was constructed,as wellas a time-saving optimal searching algorithm was designed.By the two means,the nailfold capillaries Were auto-tracked successfully.