EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Overcoming labelled data scarcity for defect classification in scanning tunnelling microscopy

Nikola L Kolev, Max Trouton, Filippo Federici Canova, Geoff Thornton, David Z Gao, Neil J Curson, Taylor J Z Stock

Machine Learning: Science and Technology · 01.04.2026 · 7 (2) · 025041

DOI: 10.1088/2632-2153/ae520e ↗

Kurz gefasst

Fachartikel in der Zeitschrift Machine Learning: Science and Technology, online erschienen am 01.04.2026. Verfasst von Nikola L Kolev und 6 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 57 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Nanolayers Research Computing, Engineering and Physical Sciences Research Council, European Research Council.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Scanning tunnelling microscopy (STM) is a powerful technique for imaging surfaces with atomic resolution, providing insight into physical and chemical processes at the level of single atoms and molecules. A regular task of STM image analysis is the identification and labelling of features of interest against a uniform background. Performing this manually is a labour-intensive task, requiring significant human effort. To reduce this burden, we propose an automated approach to the segmentation of STM images that uses both few-shot learning and unsupervised learning. Our technique offers greater flexibility compared to previous supervised methods; it removes the requirement for large manually annotated datasets and is thus easier to adapt to an unseen surface. We demonstrate the effectiveness of our approach by using it to recognise atomic features on three distinct surfaces: Si(001), Ge(001), and TiO 2 (110), including adsorbed AsH 3 molecules on the silicon and germanium surfaces. Our model exhibits potential for generalisation, and following initial training, can be adapted to unseen surfaces with as few as one additional labelled data point. This work is a significant step towards efficient and material-agnostic, automatic segmentation of STM images.

Förderung

  • Nanolayers Research Computing
  • Engineering and Physical Sciences Research Council
  • European Research Council
  • Innovate UK

Zitieren

APA

Kolev, N. L., Trouton, M., Canova, F. F., Thornton, G., Gao, D. Z., Curson, N. J., & Stock, T. J. Z. (2026). Overcoming labelled data scarcity for defect classification in scanning tunnelling microscopy. Machine Learning: Science and Technology, 7(2), 025041. https://doi.org/10.1088/2632-2153/ae520e

Vancouver

Kolev NL, Trouton M, Canova FF, Thornton G, Gao DZ, Curson NJ, et al. Overcoming labelled data scarcity for defect classification in scanning tunnelling microscopy. Machine Learning: Science and Technology. 2026;7(2):025041. doi:10.1088/2632-2153/ae520e

BibTeX
@article{Kolev2026,
  author = {Kolev, Nikola L and Trouton, Max and Canova, Filippo Federici and Thornton, Geoff and Gao, David Z and Curson, Neil J and Stock, Taylor J Z},
  title = {Overcoming labelled data scarcity for defect classification in scanning tunnelling microscopy},
  journal = {Machine Learning: Science and Technology},
  year = {2026},
  volume = {7},
  number = {2},
  pages = {025041},
  doi = {10.1088/2632-2153/ae520e}
}
Kostenlos starten →