Cross-dataset few-shot deep learning for landmark detection in biological morphometrics

Abstract

Geometric morphometrics relies on biologically homologous landmarks to quantify shape variation, yet acquiring sufficient landmark annotations remains a major obstacle to automated analysis, particularly for small, rare, or newly collected biological datasets. We propose an annotation-efficient deep learning framework for biological landmark detection that separates transferable visual representation learning from target-specific landmark adaptation. The framework first learns a general visual representation through masked autoencoding and then adapts it to a new landmark configuration using few-shot heatmap regression. This design avoids supervised episodic meta-training and allows the same pretrained encoder and adaptation procedure to be reused across datasets with different species, anatomical structures, imaging modalities, and landmark configurations. Importantly, the framework also supports target-excluded transfer, in which representation learning is performed without access to the target dataset before adaptation using only a small annotated subset. Evaluation on heterogeneous biological landmark datasets shows that the proposed framework can achieve reliable localization with limited supervision while maintaining a common architecture and training strategy across substantially different morphometric tasks. By reducing both annotation requirements and dataset-specific model engineering, the proposed framework offers a scalable route for applying deep landmark detection to small and heterogeneous biological collections.

Arthropods communities in organic orchards and adjacent hedgerows on Terceira Island (Azores, Portugal): a dataset supporting biological control research
Sophie Wallon, António Domingues, Alejandra Ros-Prieto +3
Biodiversity Data Journal27 Oct 2026Peer-reviewedCC BY 4.0Full textZoology & animal behaviourPMC
A Novel Public Dataset for Strawberry (Fragaria × Ananassa) Ripeness Detection and Comparative Evaluation of YOLO ‐Based Models
Mustafa Yurdakul, Zeynep Sena Bastug, Ali Emre Gök +1
Food Science & Nutrition1 Jul 2026Peer-reviewedCC BY7 citedPlant biologyOpenAlex
How large an effect would we have seen? Positive controls, detection floors and dataset pitfalls in scalp-EEG speech decoding
Blessing Isoyiza Adeika, Onyema Osuagwu
arXiv14 Oct 2026PreprintarXiv licenceFull textBioinformatics & computational biology NeurosciencearXiv
An annotated cucumber leaf image dataset for detection of visible damage and abnormal color symptoms in complex cultivation environments
Batyrkhan Omarov, Sayat Ibrayev, Nurbibi Imanbayeva
Frontiers in plant science2 Oct 2026Peer-reviewedCC BYFull textPlant biologyEurope PMCCrossref
A 2026 Outbreak of Locusta migratoria (Orthoptera: Acrididae) in Korea: Evidence of Partial Gregarization from Sex-Specific Morphometrics
Sumin Oh, Woong Choi, Kwang-Ho Kim +4
Insects1 Oct 2026Peer-reviewedCC BYFull textGeneral biologyEurope PMC

Cite

BibTeX RIS

BibTeX

@article{nguyen2026crossdataset,
  title = {{Cross-dataset few-shot deep learning for landmark detection in biological morphometrics}},
  author = {Hoang Ha Nguyen and Anh Tuan Giang and Nhat Quang Doan},
  journal = {Ecological Informatics},
  year = {2026},
  month = oct,
  doi = {10.1016/j.ecoinf.2026.104106},
  url = {https://doi.org/10.1016/j.ecoinf.2026.104106},
  note = {CC BY},
}

RIS

TY  - JOUR
TI  - Cross-dataset few-shot deep learning for landmark detection in biological morphometrics
AU  - Hoang Ha Nguyen
AU  - Anh Tuan Giang
AU  - Nhat Quang Doan
PY  - 2026
DA  - 2026/10/30
JO  - Ecological Informatics
DO  - 10.1016/j.ecoinf.2026.104106
UR  - https://doi.org/10.1016/j.ecoinf.2026.104106
AB  - Geometric morphometrics relies on biologically homologous landmarks to quantify shape variation, yet acquiring sufficient landmark annotations remains a major obstacle to automated analysis, particularly for small, rare, or newly collected biological datasets. We propose an annotation-efficient deep learning framework for biological landmark detection that separates transferable visual representation learning from target-specific landmark adaptation. The framework first learns a general visual representation through masked autoencoding and then adapts it to a new landmark configuration using few-shot heatmap regression. This design avoids supervised episodic meta-training and allows the same pretrained encoder and adaptation procedure to be reused across datasets with different species, anatomical structures, imaging modalities, and landmark configurations. Importantly, the framework also supports target-excluded transfer, in which representation learning is performed without access to the target dataset before adaptation using only a small annotated subset. Evaluation on heterogeneous biological landmark datasets shows that the proposed framework can achieve reliable localization with limited supervision while maintaining a common architecture and training strategy across substantially different morphometric tasks. By reducing both annotation requirements and dataset-specific model engineering, the proposed framework offers a scalable route for applying deep landmark detection to small and heterogeneous biological collections.
SN  - 1574-9541
SN  - 1878-0512
ER  -

Abstract

Related

Cite