Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: Nov 1, 2025

Tracking Rats in Operant Conditioning Chambers Using a Versatile Homemade Video Camera and DeepLabCut
08:32

Tracking Rats in Operant Conditioning Chambers Using a Versatile Homemade Video Camera and DeepLabCut

Published on: June 15, 2020

12.9K

Deep learning-based pose estimation for African ungulates in zoos.

Max Hahn-Klimroth1, Tobias Kapetanopoulos1, Jennifer Gübert2

  • 1Department of Computer Science and Mathematics Goethe University Frankfurt Germany.

Ecology and Evolution
|June 18, 2021
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Global Data From Great Ape Zoo Populations Confirm a High Prevalence of Overweight Individuals.

American journal of primatology·2026
Same author

Connection to Nature in Zoos: A Systematic Review of Empirical Research.

Zoo biology·2026
Same author

Pawsitive impact: exploring associations between pet keeping and connection to nature.

Scientific reports·2026
Same author

Behavioral Patterns of Zoo-housed Lowland (Tapirus terrestris) and Malayan tapirs (Tapirus indicus), With a Focus on Sleeping Behavior.

Zoo biology·2026
Same author

From leaves to microbes: How diet and season shape the fecal microbiome of captive Coquerel´s Sifakas (<i>Propithecus coquereli</i>).

Frontiers in microbiomes·2026
Same author

No systematic difference in obesity susceptibility between two groups of zoo primates with distinct digestive physiologies: hindgut-fermenting lemurs and foregut-fermenting colobines.

Primates; journal of primatology·2026

Automated deep learning models can now analyze wildlife behavior from video, overcoming the limitations of human observation. This new system accurately classifies animal stances, improving ecological research efficiency.

Area of Science:

  • Ecology and ethology
  • Computer vision and machine learning

Background:

  • Long-term animal behavior analysis is crucial in ecology but limited by time and cost of human observation.
  • Video recordings extend observation but manual evaluation remains time-consuming.
  • Automated evaluation using deep learning offers a solution for analyzing large video datasets efficiently.

Purpose of the Study:

  • To develop and evaluate a deep learning system for automated detection of typical stances in African ungulates.
  • To enhance system robustness through model averaging and postprocessing rules.
  • To assess the system's performance on both in-domain and out-of-domain classification tasks.

Main Methods:

  • A multistep convolutional neural network (CNN) system was designed for stance detection.
Keywords:
animal behavior statesautomated monitoringconvolutional neural networksdeep learning toolsecology of savannah animalsimage classification

More Related Videos

Decoding Natural Behavior from Neuroethological Embedding
08:00

Decoding Natural Behavior from Neuroethological Embedding

Published on: October 3, 2025

151
A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis
05:41

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

Published on: February 6, 2020

9.6K

Related Experiment Videos

Last Updated: Nov 1, 2025

Tracking Rats in Operant Conditioning Chambers Using a Versatile Homemade Video Camera and DeepLabCut
08:32

Tracking Rats in Operant Conditioning Chambers Using a Versatile Homemade Video Camera and DeepLabCut

Published on: June 15, 2020

12.9K
Decoding Natural Behavior from Neuroethological Embedding
08:00

Decoding Natural Behavior from Neuroethological Embedding

Published on: October 3, 2025

151
A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis
05:41

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

Published on: February 6, 2020

9.6K
  • Model averaging and postprocessing rules were implemented to improve robustness.
  • The system was trained and tested on video data of African ungulates in zoo enclosures.

    • Main Results:

    • The system achieved high in-domain classification accuracy (>0.92), further improved to >0.96 with postprocessing.
    • Out-of-domain classification on two unknown species yielded an average accuracy of 0.93, demonstrating robustness.
    • The developed system is publicly available for researchers to train custom models.

    Conclusions:

    • Multistep CNNs provide fast and accurate wildlife behavior classification, significantly reducing manual analysis effort.
    • Postprocessing rules effectively enable species-specific adjustments and enhance behavioral phase accuracy.
    • The system's robustness suggests applicability in diverse settings, including field studies, for various species.