A Representative Sampling Approach to Few-Shot Transfer Learning
Norsk IKT-konferanse for forskning og utdanning · 2025
DOI: 10.5324/kkyvdd97 ↗Kurz gefasst
Fachartikel in der Zeitschrift Norsk IKT-konferanse for forskning og utdanning, 2025. Verfasst von Rusith Chamara Hathurusinghe Dewage und 4 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref). Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Accurate plant growth estimation is essential in controlledenvironment agriculture (CEA), where crop monitoring, yield prediction,and environmental optimization all rely on reliable assessments. However,this task remains challenging due to the limited availability of labeleddata. In this study, we propose a few-shot transfer learning frameworkfor estimating cucumber plant growth, using a multi-stage approach.Our method combines deep feature extraction, representative image se-lection, and regression to accurately predict cucumber plant height usinglimited samples. By identifying key representative images, our approachreduces data requirements while maintaining comparable predictive per-formance. Our best model achieved a mean absolute error (MAE) of20 mm with an R2 value of 0.68 using six representative samples. Ex-perimental results demonstrate that systematic representative samplingcombined with regression significantly improves predictive performanceand stability in few-shot transfer learning scenarios, thereby minimizingdata collection efforts and reducing computational costs. Moreover, theintegration of clustering enhances explainability, making the approachmore transparent and suitable for practical deployment.
Zitieren
APA
Dewage, R. C. H., Ullah, H., Siddiqui, M. S., Islam, R., & Machot, F. (2025). A Representative Sampling Approach to Few-Shot Transfer Learning. Norsk IKT-konferanse for forskning og utdanning. https://doi.org/10.5324/kkyvdd97
Vancouver
Dewage RCH, Ullah H, Siddiqui MS, Islam R, Machot F. A Representative Sampling Approach to Few-Shot Transfer Learning. Norsk IKT-konferanse for forskning og utdanning. 2025. doi:10.5324/kkyvdd97
BibTeX
@article{Dewage2025,
author = {Dewage, Rusith Chamara Hathurusinghe and Ullah, Habib and Siddiqui, Muhammad Salman and Islam, Rakibul and Machot, Fadi-Al},
title = {A Representative Sampling Approach to Few-Shot Transfer Learning},
journal = {Norsk IKT-konferanse for forskning og utdanning},
year = {2025},
doi = {10.5324/kkyvdd97}
}
EUVIMED