EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Integrating machine learning, deep learning, and image analysis for seed species classification

Jonathan Ashworth iD, Eloise Barrett iD, Rebonto Haque iD, Ben Wildey, Anthony Hall iD, Felix Shaw, Jose J De Vega iD

Applications in Plant Sciences · 2026-09 · 14 (5)

DOI: 10.1002/aps3.70072 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Applications in Plant Sciences, online erschienen am 23.07.2026. Verfasst von Jonathan Ashworth und 6 weiteren Autor:innen. Beteiligte Einrichtung: Earlham Institute Norwich Research Park Norwich NR4 7UZ United Kingdom. Die Arbeit nennt 39 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Biotechnology and Biological Sciences Research Council, Natural Environment Research Council.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Premise The growing demand for wildflower seeds in ecological restoration requires reliable species identification, yet current market products often contain heterogeneous species. As seed identification is labor‐intensive and requires advanced botanical knowledge, we evaluated multiple segmentation and classification approaches to determine which combination performed better for identifying species and quantifying their frequencies under real‐life constraints.

Methods

We generated seed images from common wildflower species using a flatbed scanner. Artificial intelligence (AI) and non‐AI segmentation pipelines were compared. We evaluated over 20 classifiers on tabulated features and convolutional neural networks (CNNs) on single seed image inputs. Open‐set conditions were simulated using “mock seed mixes” with previously unseen taxa using decision thresholds.

Results

With tabulated data, the XGBoost and Multi‐layer Perceptron (MLP) models exceeded F1 > 0.96, while AutoGluon's CNN and ensemble models reached F1 > 0.97. ResNet‐50 achieved an F1 score over 0.99 on single‐seed images obtained with Cellpose segmentation. Random forest achieved the highest accuracy (0.92) with unseen species in open‐set classification.

Discussion

Deep learning segmentation substantially enhanced CNN accuracy, yet tabulated‐feature machine learning models remained competitive and efficient. Threshold‐based rejection enabled robust open‐set classification, with random forest outperforming deeper models in rejecting unseen seeds. No single method was universally optimal; the best strategy depends on whether the task involves closed‐set or open‐set classification and on trade‐offs between accuracy and computational cost.

Schlagwörter

Stichwörter der Autor:innen

habitat restorationimage analysismachine learningseed provenancespecies classificationwildflowers

Autor:innen und Institutionen

  • Earlham Institute Norwich Research Park Norwich NR4 7UZ United Kingdom.

ORCID-Profile: Jonathan Ashworth, Eloise Barrett, Rebonto Haque, Anthony Hall, Jose J De Vega

Förderung

  • Biotechnology and Biological Sciences Research Council
  • Natural Environment Research Council

Zitieren

APA

Ashworth, J., Barrett, E., Haque, R., Wildey, B., Hall, A., Shaw, F., & De Vega, J. J. (2026). Integrating machine learning, deep learning, and image analysis for seed species classification. Applications in Plant Sciences, 14(5). https://doi.org/10.1002/aps3.70072

Vancouver

Ashworth J, Barrett E, Haque R, Wildey B, Hall A, Shaw F, et al. Integrating machine learning, deep learning, and image analysis for seed species classification. Applications in Plant Sciences. 2026;14(5). doi:10.1002/aps3.70072

BibTeX
@article{Ashworth2026,
  author = {Ashworth, Jonathan and Barrett, Eloise and Haque, Rebonto and Wildey, Ben and Hall, Anthony and Shaw, Felix and De Vega, Jose J},
  title = {Integrating machine learning, deep learning, and image analysis for seed species classification},
  journal = {Applications in Plant Sciences},
  year = {2026},
  volume = {14},
  number = {5},
  doi = {10.1002/aps3.70072}
}
Kostenlos starten →