Machine Learning Implementation in live-cell tracking

While quantitative biology has gradually become the major trend of biology, researchers have put their eyes on analysis tools that can give them more quantitative read-out of the biological system. Cell biologists are researchers that have started to use quantitative image analysis tools at the earliest time, and by using these tools, information such as the dynamics of certain molecules of interest can be quantified rather easily. Among all the approaches, live-cell cell tracking has become the most informative but also the most challenging one. To date, most of the cell-tracking scripts are semi-automatic of which parameters have to be heavily tuned as well as observational supervision. To improve this method and make it more automatic is thus a meaningful problem. In this project, we are hoping to implement machine learning into the tracking dataset and look for potential improvement of the original tracking method.