EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Multi-scale prototype convolutional network for few-shot semantic segmentation

Ding Xu iD, Shun Yu, Jingxuan Zhou, Fusen Guo iD, Lin Li, Jishizhan Chen iD

PLOS ONE · 15.04.2025 · 20 (4) · e0319905

DOI: 10.1371/journal.pone.0319905 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS ONE, online erschienen am 15.04.2025. Verfasst von Ding Xu und 5 weiteren Autor:innen. Beteiligte Einrichtung: Computer Science Department, Harbin Institute of Technology, Harbin, China. Hauptthemen laut MEDLINE: Semantics, Neural Networks, Computer. Die Arbeit wurde bisher 13-mal zitiert (Crossref) und nennt 47 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Few-shot semantic segmentation aims to accurately segment objects from a limited amount of annotated data, a task complicated by intra-class variations and prototype representation challenges. To address these issues, we propose the Multi-Scale Prototype Convolutional Network (MPCN). Our approach introduces a Prior Mask Generation (PMG) module, which employs dynamic kernels of varying sizes to capture multi-scale object features. This enhances the interaction between support and query features, thereby improving segmentation accuracy. Additionally, we present a Multi-Scale Prototype Extraction (MPE) module to overcome the limitations of MAP (Mean Average Precision). By augmenting support set features, assessing spatial importance, and utilizing multi-scale downsampling, we obtain a more accurate prototype set. Extensive experiments conducted on the PASCAL- 5 i and COCO- 2 0 i datasets demonstrate that our method achieves superior performance in both 1-shot and 5-shot settings.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

SemanticsNeural Networks, ComputerAlgorithmsHumans

Autor:innen und Institutionen

  • Computer Science Department, Harbin Institute of Technology, Harbin, China.
  • School of Systems and Computing, University of New South Wales, Canberra, Australia.
  • Faculty of Science and Engineering, Southern Cross University, Gold Coast, Australia.
  • Mechanical Engineering, University College London, London, United Kingdom.

ORCID-Profile: Ding Xu, Fusen Guo, Jishizhan Chen

Zitieren

APA

Xu, D., Yu, S., Zhou, J., Guo, F., Li, L., & Chen, J. (2025). Multi-scale prototype convolutional network for few-shot semantic segmentation. PLOS ONE, 20(4), e0319905. https://doi.org/10.1371/journal.pone.0319905

Vancouver

Xu D, Yu S, Zhou J, Guo F, Li L, Chen J. Multi-scale prototype convolutional network for few-shot semantic segmentation. PLOS ONE. 2025;20(4):e0319905. doi:10.1371/journal.pone.0319905

BibTeX
@article{Xu2025,
  author = {Xu, Ding and Yu, Shun and Zhou, Jingxuan and Guo, Fusen and Li, Lin and Chen, Jishizhan},
  title = {Multi-scale prototype convolutional network for few-shot semantic segmentation},
  journal = {PLOS ONE},
  year = {2025},
  volume = {20},
  number = {4},
  pages = {e0319905},
  doi = {10.1371/journal.pone.0319905}
}
Kostenlos starten →