Article Dans Une Revue Journal of Machine Learning for Biomedical Imaging Année : 2025

Sketchpose: Learning to Segment Cells with Partial Annotations

Résumé

The most popular networks used for cell segmentation (e.g. Cellpose, Stardist, HoverNet,...) rely on a prediction of a distance map. It yields unprecedented accuracy but hinges on fully annotated datasets. This is a serious limitation to generate training sets and perform transfer learning. In this paper, we propose a method that still relies on the distance map and handles partially annotated objects. We evaluate the performance of the proposed approach in the contexts of frugal learning, transfer learning and regular learning on regular databases. Our experiments show that it can lead to substantial savings in time and resources without sacrificing segmentation quality. The proposed algorithm is embedded in a user-friendly Napari plugin.

Fichier principal
Vignette du fichier
sketchpose1.pdf (34.23 Mo) Télécharger le fichier
Origine Fichiers éditeurs autorisés sur une archive ouverte
Licence
DOI

Cite 10.5281/zenodo.6497715 Jeu de données Scherr, T., Seiffarth, J., Wollenhaupt, B., Neumann, O., Schilling, M. P., Kohlheyer, D., Scharr, H., Nöh, K., & Mikut, R. (2022). microbeSEG dataset (Version 1.0) [Data set]. Zenodo. https://doi.org/10.5281/ZENODO.6497715

MicrobeSeg (dataset)
DOI

Cite 10.5281/zenodo.7221152 Article Scherr, T., Seiffarth, J., Wollenhaupt, B., Neumann, O., Kohlheyer, D., Scharr, H., Nöh, K., & Mikut, R. (2022). microbeSEG models (Version 1.0). Zenodo. https://doi.org/10.5281/ZENODO.7221152

MicrobeSeg (dataset)

Dates et versions

hal-04330824 , version 1 (08-12-2023)
hal-04330824 , version 2 (25-08-2025)

Licence

Identifiants

Citer

Clément Cazorla, Nathanaël Munier, Renaud Morin, Pierre Weiss. Sketchpose: Learning to Segment Cells with Partial Annotations. Journal of Machine Learning for Biomedical Imaging, 2025, ⟨10.59275/j.melba.2025-f7b3⟩. ⟨hal-04330824v2⟩
1396 Consultations
461 Téléchargements

Altmetric

Partager

  • More