LiveCellTrack: Label-free Cell Tracking in High Confluence 2D Cell Cultures --Preview

Published: 5 March 2025| Version 1 | DOI: 10.17632/cgwcpz34mr.1
Contributors:
,
, Maria Caroprese, Jasmine Trigg, Rickard Sjogren, Mark Owen, Johan Trygg, Andreas Dengel, Sheraz Ahmed

Description

Cell tracking - the ability to quantify the actions of a single cell through time - provides researchers with highly detailed information on cell behaviour and function. Image-based in vitro assays – which are used to monitor cells over several days - display dynamically changing conditions with cells gradually altering morphology as they move around the field of view and proliferate, eventually covering the full field of view. Tracking cells under these challenging conditions requires firstly, robust detection of all cells in the image, and secondly, accurate connection of individual cells from frame to frame. Here we introduce LiveCellTrack dataset which provide image sequences showing cells under a range of biological conditions including high confluence, manually annotated to a high standard wherein detected cells are linked with unique ID numbers through each image sequence. In addition, we demonstrate an analysis workflow which identifies and tracks cells through time with high accuracy relative to the ground truth manually annotated data. This is a preview (subset) of the original dataset

Files

Institutions

Sartorius AG, Deutsches Forschungszentrum fur Kunstliche Intelligenz GmbH

Categories

Machine Learning, Biological Database, Cellular Imaging, Phase Contrast Microscopy

Licence