Abstract
Single-cell quantitative time-lapse microscopy is a valuable tool for directly observing the dynamic activities and decisions of individual stem cells within a population. Successful application of time-lapse microscopy requires automated segmentation of thousands of individual cells to track their features frame-by-frame over time. However, this can be challenging due to unpredictable cell movement, changes in cell shape, and mitotic activity. Such limitations are particularly notable in imaging approaches such as label-free phase-contrast microscopy, which is widely available and non-toxic for long-term time-lapse imaging. Our team has previously developed a deep-learning model and software, DeepSea, to overcome the challenges of single-cell phase-contrast microscopy through automation of segmentation and tracking. In this chapter, we present how to install and use DeepSea to segment and track individual cells in time-lapse phase-contrast images.