Low-Frame-Rate Cell Tracking: Unmet Needs and Future Directions

Deep learning has significantly improved cell tracking, but most methods rely on high-frame-rate imaging, which can cause photobleaching and phototoxicity. Low-frame-rate imaging mitigates these effects, enabling long-term observations while introducing new tracking challenges. This work presents a novel dataset for low-frame-rate cell tracking, featuring long-term image sequences with ground-truth annotations for cell identification, mitosis, and movement. We sample well-established tracking methods and reveal their limitations in handling sparse temporal data. Our analysis highlights key challenges, including increased cell displacement, missed mitosis events, and tracking ambiguities. Moreover, we discuss how to set up an objective framework for cell tracking and future directions to develop robust tracking methods tailored to low-frame-rate conditions.