EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Animal Species Recognition with Deep Convolutional Neural Networks from Ecological Camera Trap Images

Sazida Binta Islam iD, Damian Valles iD, Toby J Hibbitts, Wade A Ryberg iD, Danielle K Walkup, Michael R J Forstner

Animals · 02.05.2023 · 13 (9) · 1526

DOI: 10.3390/ani13091526 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Animals, online erschienen am 02.05.2023. Verfasst von Sazida Binta Islam und 5 weiteren Autor:innen. Beteiligte Einrichtung: Ingram School of Engineering, Texas State University, San Marcos, TX 78666, USA. Die Arbeit wurde bisher 70-mal zitiert (Crossref) und nennt 65 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: U.S. Fish and Wildlife Service, Bastrop Utilities HCP, Texas A&M Natural Resources Institute.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate identification of animal species is necessary to understand biodiversity richness, monitor endangered species, and study the impact of climate change on species distribution within a specific region. Camera traps represent a passive monitoring technique that generates millions of ecological images. The vast numbers of images drive automated ecological analysis as essential, given that manual assessment of large datasets is laborious, time-consuming, and expensive. Deep learning networks have been advanced in the last few years to solve object and species identification tasks in the computer vision domain, providing state-of-the-art results. In our work, we trained and tested machine learning models to classify three animal groups (snakes, lizards, and toads) from camera trap images. We experimented with two pretrained models, VGG16 and ResNet50, and a self-trained convolutional neural network (CNN-1) with varying CNN layers and augmentation parameters. For multiclassification, CNN-1 achieved 72% accuracy, whereas VGG16 reached 87%, and ResNet50 attained 86% accuracy. These results demonstrate that the transfer learning approach outperforms the self-trained model performance. The models showed promising results in identifying species, especially those with challenging body sizes and vegetation.

Schlagwörter

Stichwörter der Autor:innen

camera trapconvolutional neural networkdeep learningendangered speciesimage augmentationimage classificationlizardmachine learningsnaketoad

Autor:innen und Institutionen

  • Ingram School of Engineering, Texas State University, San Marcos, TX 78666, USA.
  • Natural Resources Institute, Texas A&M University, College Station, TX 77843, USA.
  • Biodiversity Research and Teaching Collections, Texas A&M University, College Station, TX 77843, USA.
  • Department of Biology, Texas State University, San Marcos, TX 78666, USA.

ORCID-Profile: Sazida Binta Islam, Damian Valles, Wade A Ryberg

Förderung

  • U.S. Fish and Wildlife Service
  • Bastrop Utilities HCP
  • Texas A&M Natural Resources Institute
  • Texas Comptroller of Public Accounts
  • Texas Research Incentive Program

Zitieren

APA

Binta Islam, S., Valles, D., Hibbitts, T. J., Ryberg, W. A., Walkup, D. K., & Forstner, M. R. J. (2023). Animal Species Recognition with Deep Convolutional Neural Networks from Ecological Camera Trap Images. Animals, 13(9), 1526. https://doi.org/10.3390/ani13091526

Vancouver

Binta Islam S, Valles D, Hibbitts TJ, Ryberg WA, Walkup DK, Forstner MRJ. Animal Species Recognition with Deep Convolutional Neural Networks from Ecological Camera Trap Images. Animals. 2023;13(9):1526. doi:10.3390/ani13091526

BibTeX
@article{BintaIslam2023,
  author = {Binta Islam, Sazida and Valles, Damian and Hibbitts, Toby J and Ryberg, Wade A and Walkup, Danielle K and Forstner, Michael R J},
  title = {Animal Species Recognition with Deep Convolutional Neural Networks from Ecological Camera Trap Images},
  journal = {Animals},
  year = {2023},
  volume = {13},
  number = {9},
  pages = {1526},
  doi = {10.3390/ani13091526}
}
Kostenlos starten →