CellBinDB: a large-scale multimodal annotated dataset for cell segmentation with benchmarking of universal models
GigaScience · 06.01.2025 · 14
DOI: 10.1093/gigascience/giaf069 ↗Kurz gefasst
Validierungsstudie in der Zeitschrift GigaScience, online erschienen am 24.06.2025. Verfasst von Can Shi und 12 weiteren Autor:innen. Beteiligte Einrichtung: BGI Research, Shenzhen 518083, China. Hauptthemen laut MEDLINE: Cells, Deep Learning, Staining and Labeling, Tissues. Die Arbeit wurde bisher 8-mal zitiert (Crossref) und nennt 61 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Key Research and Development Program of China.
Abstract (Originalsprache)
In recent years, cell segmentation techniques have played a critical role in the analysis of biological images, especially for quantitative studies. Deep learning–based cell segmentation models have demonstrated remarkable performance in segmenting cell and nucleus boundaries, but they are typically tailored to specific modalities or require manual tuning of hyperparameters, limiting their generalizability to unseen data. Comprehensive datasets that support both the training of universal models and the evaluation of various segmentation techniques are essential for overcoming these limitations and promoting the development of more versatile cell segmentation solutions. Here, we present CellBinDB, a large-scale multimodal annotated dataset established for these purposes. CellBinDB contains more than 1,000 annotated images, each labeled to identify the boundaries of cells or nuclei, including 4′,6-diamidino-2-phenylindole, single-stranded DNA, hematoxylin and eosin, and multiplex immunofluorescence staining, covering over 30 normal and diseased tissue types from human and mouse samples. Based on CellBinDB, we benchmarked 8 state-of-the-art and widely used cell segmentation technologies/methods, and our further analysis reveals that complex cell shapes reduce segmentation accuracy while higher image gradients improve boundary detection, offering insights for refining segmentation strategies across diverse imaging scenarios.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- BGI Research, Shenzhen 518083, China.
- School of Artificial Intelligence, University of China Academy of Science, Beijing 100049, China.
ORCID-Profile: Can Shi, Jinghong Fan, Zhonghan Deng, Huanlin Liu, Qiang Kang, Yumei Li, Jing Guo, Jingwen Wang, Jinjiang Gong, Sha Liao, Ao Chen, Ying Zhang, Mei Li
Förderung
- National Key Research and Development Program of China
Zitieren
APA
Shi, C., Fan, J., Deng, Z., Liu, H., Kang, Q., Li, Y., Guo, J., Wang, J., Gong, J., Liao, S., Chen, A., Zhang, Y., & Li, M. (2025). CellBinDB: a large-scale multimodal annotated dataset for cell segmentation with benchmarking of universal models. GigaScience, 14. https://doi.org/10.1093/gigascience/giaf069
Vancouver
Shi C, Fan J, Deng Z, Liu H, Kang Q, Li Y, et al. CellBinDB: a large-scale multimodal annotated dataset for cell segmentation with benchmarking of universal models. GigaScience. 2025;14. doi:10.1093/gigascience/giaf069
BibTeX
@article{Shi2025,
author = {Shi, Can and Fan, Jinghong and Deng, Zhonghan and Liu, Huanlin and Kang, Qiang and Li, Yumei and Guo, Jing and Wang, Jingwen and Gong, Jinjiang and Liao, Sha and Chen, Ao and Zhang, Ying and Li, Mei},
title = {CellBinDB: a large-scale multimodal annotated dataset for cell segmentation with benchmarking of universal models},
journal = {GigaScience},
year = {2025},
volume = {14},
doi = {10.1093/gigascience/giaf069}
}
EUVIMED