Illustration of key steps in the segmentation and tracking of cell-pair rotation timelapses. Membrane probability: output of the Ilastik autocontext workflow trained to classify pixels as membrane/non-membrane. The probability is an input for the Ilastik multicut workflow and the resulting skeletonized segmentation of cell boundaries is manually refined in Tissue analyser. Cells are then tracked over time to give an evolving topological representation of the tissue in general, and the cell-pair rotation in particular.
Eva L. Kozak, Jerónimo R. Miranda-Rodríguez, Augusto Borges, Kai Dierkes, Alessandro Mineo, Filipe Pinto-Teixeira, Oriol Viader-Llargués, Jérôme Solon, Osvaldo Chara, and Hernán López-Schier
Development
2023. 150:None-None; doi: 10.1242/dev.200975