EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Improving the accuracy of automated labeling of specimen images datasets via a confidence-based process

Quentin Bateux iD, Jonathan Koss, Patrick W Sweeney, Erika Edwards, Nelson Rios, Aaron M Dollar

PLOS Computational Biology · 2025

DOI: 10.1371/journal.pcbi.1013650 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS Computational Biology, 2025. Verfasst von Quentin Bateux und 5 weiteren Autor:innen. Beteiligte Einrichtung: Institute for Biospheric Studies, Yale University, New Haven, Connecticut, United States of America. Hauptthemen laut MEDLINE: Image Processing, Computer-Assisted. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 82 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Yale Institute for Biospheric Studies, Institute for Biospheric Studies, Yale University.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The digitization of natural history collections over the past three decades has unlocked a treasure trove of specimen imagery and metadata. There is great interest in making this data more useful by further labeling it with additional trait data, and modern “deep learning” machine learning techniques utilizing convolutional neural nets (CNNs) and similar networks show particular promise to reduce the amount of required manual labeling by human experts, making the process much faster and less expensive. However, in most cases, the accuracy of these approaches is too low for reliable utilization of the automatic labeling, typically in the range of 80-85% accuracy. In this paper, we present and validate an approach that can greatly improve this accuracy, essentially by examining the “confidence” that the network has in the generated label as well as utilizing a user-defined threshold to reject labels that fall below a chosen level. We demonstrate that a naive model that produced 86% initial accuracy can achieve improved performance - over 95% accuracy (rejecting about 40% of the labels) or over 99% accuracy (rejecting about 65%) by selecting higher confidence thresholds. This gives flexibility to adapt existing models to the statistical requirements of various types of research and has the potential to move these automatic labeling approaches from being unusably inaccurate to being an invaluable new tool. After validating the approach in a number of ways, we annotate the reproductive state of a large dataset of over 600,000 herbarium specimens. The analysis of the results points at under-investigated correlations as well as general alignment with known trends. By sharing this new dataset alongside this work, we want to allow biologists to gather insights for their own research questions, at their chosen point of accuracy/coverage trade-off.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Image Processing, Computer-AssistedHumansNeural Networks, ComputerComputational BiologyMachine LearningAlgorithmsDatabases, FactualDeep Learning

Autor:innen und Institutionen

  • Institute for Biospheric Studies, Yale University, New Haven, Connecticut, United States of America.
  • Department of Electrical Engineering, Yale University, New Haven, Connecticut, United States of America.
  • Division of Botany at the Yale Peabody Museum, Yale University, New Haven, Connecticut, United States of America.
  • Department of Ecology and Evolutionary Biology, Yale University, New Haven, Connecticut, United States of America.
  • Peabody Museum and Yale School of Forestry and Environmental Studies, Yale University, New Haven, Connecticut, United States of America.
  • Department of Mechanical Engineering and Materials Science, Yale University, New Haven, Connecticut, United States of America.

ORCID-Profile: Quentin Bateux

Förderung

  • Yale Institute for Biospheric Studies
  • Institute for Biospheric Studies, Yale University

Zitieren

APA

Bateux, Q., Koss, J., Sweeney, P. W., Edwards, E., Rios, N., & Dollar, A. M. (2025). Improving the accuracy of automated labeling of specimen images datasets via a confidence-based process. PLOS Computational Biology. https://doi.org/10.1371/journal.pcbi.1013650

Vancouver

Bateux Q, Koss J, Sweeney PW, Edwards E, Rios N, Dollar AM. Improving the accuracy of automated labeling of specimen images datasets via a confidence-based process. PLOS Computational Biology. 2025. doi:10.1371/journal.pcbi.1013650

BibTeX
@article{Bateux2025,
  author = {Bateux, Quentin and Koss, Jonathan and Sweeney, Patrick W and Edwards, Erika and Rios, Nelson and Dollar, Aaron M},
  title = {Improving the accuracy of automated labeling of specimen images datasets via a confidence-based process},
  journal = {PLOS Computational Biology},
  year = {2025},
  doi = {10.1371/journal.pcbi.1013650}
}
Kostenlos starten →