Evaluation of cell segmentation methods without reference segmentations
Authors:
- 15 May 2023
- journal article
- research article
- Published by American Society for Cell Biology (ASCB) in Molecular Biology of the Cell
- Vol. 34 (6) , ar50
Abstract
We developed metrics and an open source tool to evaluate the cell segmentation quality of multichannel images without a human-labeled reference. Deep learning–based methods performed the best, but postprocessing to ensure proper matching of cell and nuclear masks dramatically improved performance.Keywords
All Related Versions
This publication has 36 references indexed in Scilit:
- Active Contour-Based Cell Segmentation During Freezing and Its Application in CryopreservationIEEE Transactions on Biomedical Engineering, 2014
- Accurate cell segmentation in microscopy images using membrane patternsBioinformatics, 2014
- Multiplexed ion beam imaging of human breast tumorsNature Medicine, 2014
- A benchmark for comparison of cell tracking algorithmsBioinformatics, 2014
- Cell migration or cytokinesis and proliferation? – Revisiting the “go or grow” hypothesis in cancer cells in vitroExperimental Cell Research, 2013
- CellSegm - a MATLAB toolbox for high-throughput 3D cell segmentationSource Code for Biology and Medicine, 2013
- Highly multiplexed single-cell analysis of formalin-fixed, paraffin-embedded cancer tissueProceedings of the National Academy of Sciences, 2013
- Using CellX to Quantify Intracellular EventsCurrent Protocols in Molecular Biology, 2013
- Improved structure, function and compatibility for CellProfiler: modular high-throughput image analysis softwareBioinformatics, 2011
- Characterization of Biological Processes through Automated Image AnalysisAnnual Review of Biomedical Engineering, 2010