EUVIMEDOpdag EUVIMED
Crossref

Herbariograph : a deep‐learning tool to classify specimen images

Fabio Andrés Ávila iD, John Y Park iD, Leanna Feder iD, Damon P Little iD

New Phytologist · 2026-07 · 251 (2) · 856-867

DOI: 10.1111/nph.70447 ↗

At a glance

Journal article in New Phytologist, published online on 2025-08-11. Written by Fabio Andrés Ávila and 3 co-authors. Institution: New York Botanical Garden, New York City, NY, 10458, USA. Major MEDLINE topics: Deep Learning, Image Processing, Computer-Assisted, Plants. Cited 4 times so far (Crossref); 36 references. Funded by: National Science Foundation.

AbstractSubjectsBibliografiske oplysningerCiteRelated studies

Abstract (original language)

Herbarium specimens are the physical evidence of plant diversity world‐wide and a rich source of research data. With c. 402 M items, distributed in 3.9 k registered herbaria, in 182 countries, it is estimated that c. 12.8% have associated digital images. Despite this low proportion, information in these images facilitates numerous avenues of research aimed at understanding the hidden complexity and variation of plants. The first step in harvesting this information is to sort images by the types of items represented within them. To address this challenge, we present Herbariograph , a new open image dataset and deep‐learning model designed to automatically recognize all the image types commonly stored in collection databases. The dataset consists of 17 image categories with 12 288 images per category gathered from 43 institutions. A Convolutional Neural Network was trained on the Herbariograph dataset. The trained model produces a test macro F 1 score of 0.9611. Herbariograph will help to automate specimen image processing to make targeted dataset creation faster, better, and more accessible.

Subjects

MeSH terms (MEDLINE)

Deep LearningImage Processing, Computer-AssistedPlantsClassification AlgorithmsConvolutional Neural NetworksDatabases, Factual

Author keywords

Artificial Intelligencebiodiversityherbarium specimenimage classificationimagesmachine learning

Authors and institutions

  • New York Botanical Garden, New York City, NY, 10458, USA.
  • The Graduate Center, City University of New York, New York City, NY, 10017, USA.
  • USDA Scientific Computing Initiative (SciNet)/Oak Ridge Institute for Science and Education (ORISE), Oak Ridge, TN, 37830, USA.

ORCID profiles: Fabio Andrés Ávila, John Y Park, Leanna Feder, Damon P Little

Funding

  • National Science Foundation

Cite

APA

Ávila, F. A., Park, J. Y., Feder, L., & Little, D. P. (2026). Herbariograph : a deep‐learning tool to classify specimen images. New Phytologist, 251(2), 856-867. https://doi.org/10.1111/nph.70447

Vancouver

Ávila FA, Park JY, Feder L, Little DP. Herbariograph : a deep‐learning tool to classify specimen images. New Phytologist. 2026;251(2):856-867. doi:10.1111/nph.70447

BibTeX
@article{vila2026,
  author = {Ávila, Fabio Andrés and Park, John Y and Feder, Leanna and Little, Damon P},
  title = {Herbariograph : a deep‐learning tool to classify specimen images},
  journal = {New Phytologist},
  year = {2026},
  volume = {251},
  number = {2},
  pages = {856--867},
  doi = {10.1111/nph.70447}
}
Start for free →