EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

DINOSim : Training‐Free Semantic Segmentation on Microscopy Images

Aitor González-Marfil iD, Estibaliz Gómez-de-Mariscal iD, Ignacio Arganda-Carreras iD

Microscopy Research and Technique · 28.10.2026

DOI: 10.1002/jemt.70194 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Microscopy Research and Technique, online erschienen am 28.10.2026. Verfasst von Aitor González-Marfil und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Computer Science and Artificial Intelligence, University of the Basque Country (UPV/EHU), San Sebastián, Spain. Die Arbeit nennt 37 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Ministerio de Ciencia, Innovación y Universidades.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

We present DINOSim, a training‐free method for segmenting objects in microscopy images without the need for large annotated datasets or additional model training. DINOSim builds on the pretrained DINOv2 image encoder, which captures semantic information from images. By comparing the encoder's features of image patches to those of a user‐selected reference, DINOSim generates pseudo‐labels that guide semantic segmentation. Subsequently, a k ‐nearest neighbors framework is used to refine predictions across new images. Our experiments show that DINOSim can effectively segment previously unseen objects in diverse microscopy datasets, offering performance comparable to supervised approaches while avoiding the need for costly manual labeling. We also investigate how different choices of user prompt selection and model size affect accuracy and generalization. To make the method widely accessible, we provide an open‐source napari plugin ( github.com/AAitorG/napari‐DINOSim ), enabling researchers to apply DINOSim to their own data easily. Overall, DINOSim offers a fast, flexible and practical solution for bioimage analysis, particularly valuable in resource‐constrained settings.

Schlagwörter

Stichwörter der Autor:innen

foundation modelsimage segmentationtraining‐free

Autor:innen und Institutionen

  • Department of Computer Science and Artificial Intelligence, University of the Basque Country (UPV/EHU), San Sebastián, Spain.
  • Donostia International Physics Center (DIPC), San Sebastián, Spain.
  • António Xavier Institute of Chemical and Biological Technology, Universidade Nova de Lisboa, Oeiras, Portugal.
  • NOVA Institute for Medical Systems Biology, Association, NOVA IMSB, Paço de Arcos, Portugal.
  • iNOVA4Health, Programme in Translational Medicine, Lisboa, Portugal.
  • Biofisika Institute (CSIC-UPV/EHU), Leioa, Spain.

ORCID-Profile: Aitor González-Marfil, Estibaliz Gómez-de-Mariscal, Ignacio Arganda-Carreras

Förderung

  • Ministerio de Ciencia, Innovación y Universidades

Zitieren

APA

González-Marfil, A., Gómez-de-Mariscal, E., & Arganda-Carreras, I. (2026). DINOSim : Training‐Free Semantic Segmentation on Microscopy Images. Microscopy Research and Technique. https://doi.org/10.1002/jemt.70194

Vancouver

González-Marfil A, Gómez-de-Mariscal E, Arganda-Carreras I. DINOSim : Training‐Free Semantic Segmentation on Microscopy Images. Microscopy Research and Technique. 2026. doi:10.1002/jemt.70194

BibTeX
@article{GonzlezMarfil2026,
  author = {González-Marfil, Aitor and Gómez-de-Mariscal, Estibaliz and Arganda-Carreras, Ignacio},
  title = {DINOSim : Training‐Free Semantic Segmentation on Microscopy Images},
  journal = {Microscopy Research and Technique},
  year = {2026},
  doi = {10.1002/jemt.70194}
}
Kostenlos starten →