Marganski William A, Dembo Micah, Wang Yu-Li
Department of Biomedical Engineering, Boston University, Boston, Massachusetts 02215, USA.
Methods Enzymol. 2003;361:197-211. doi: 10.1016/s0076-6879(03)61012-8.
The optical flow algorithm presented here is a robust method that rapidly yields a high-density field of substrate displacement vectors based on two optical images. We found that one of the limiting factors, at least for inexperienced experimentalists, is the consistency of focusing or the drift in microscope focus. However, with properly collected images the standard error of the measurement was estimated to be on the order of +/- 0.10 pixels. Finally, although the discussion has been focused on the displacement of flexible substrata, a similar method should be applicable for detecting movements on other types of images, as long as the movement involves a certain degree of local coordination.
本文提出的光流算法是一种稳健的方法,它基于两幅光学图像快速生成底物位移矢量的高密度场。我们发现,至少对于缺乏经验的实验人员来说,限制因素之一是聚焦的一致性或显微镜焦点的漂移。然而,通过正确采集的图像,测量的标准误差估计在±0.10像素左右。最后,尽管讨论主要集中在柔性底物的位移上,但只要运动涉及一定程度的局部协调性,类似的方法应该适用于检测其他类型图像上的运动。