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 Concept Videos

Methods of Classification and Identification01:28

Methods of Classification and Identification

Bacterial identification relies on a diverse array of techniques to classify and understand microorganisms, each tailored to uncover specific characteristics. Traditional morphological approaches, while still valuable, are limited for closely related or structurally simple organisms. Modern methods integrate biochemical, serological, genetic, and advanced molecular tools to achieve greater accuracy.Morphological and Biochemical TechniquesMorphological characteristics, such as cell shape and...

You might also read

Related Articles

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

Sort by
Same author

DJAN: Deep Joint Adaptation Network for Wildlife Image Recognition.

Animals : an open access journal from MDPI·2023
Same author

A Lightweight Automatic Wildlife Recognition Model Design Method Mitigating Shortcut Learning.

Animals : an open access journal from MDPI·2023
Same author

Identifying Animals in Camera Trap Images via Neural Architecture Search.

Computational intelligence and neuroscience·2022
See all related articles

Related Experiment Video

Updated: Jul 15, 2026

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment
08:25

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Published on: May 7, 2019

9.1K

Domain-Aware Neural Architecture Search for Classifying Animals in Camera Trap Images.

Liang Jia1,2, Ye Tian1, Junguo Zhang1

  • 1School of Technology, Beijing Forestry University, Beijing 100083, China.

Animals : an Open Access Journal From MDPI
|February 25, 2022
PubMed
Summary

Researchers can now automatically design custom neural networks for wildlife camera traps using a novel search method. This approach optimizes networks for resource-limited edge devices, enabling cost-effective wildlife surveillance.

Keywords:
camera trap imagesconvolutional neural networkneural architecture search

More Related Videos

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

9.5K
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

12.7K

Related Experiment Videos

Last Updated: Jul 15, 2026

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment
08:25

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Published on: May 7, 2019

9.1K
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

9.5K
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

12.7K

Area of Science:

  • Ecology
  • Computer Science
  • Artificial Intelligence

Background:

  • Camera traps generate vast wildlife image datasets requiring classification.
  • Automated image classification using convolutional neural networks (CNNs) on edge devices is feasible but challenging due to device heterogeneity and resource limitations.
  • Existing neural architecture search (NAS) methods are often ill-suited for camera trap data due to differences in data distribution and image aspect ratios compared to benchmark datasets.

Purpose of the Study:

  • To develop a novel neural architecture search (NAS) method tailored for optimizing CNNs directly on camera trap images.
  • To address the limitations of existing NAS methods by accounting for the unique characteristics of camera trap data, such as lower resolutions and maintained aspect ratios.
  • To generate lightweight and efficient CNNs suitable for deployment on resource-constrained edge devices for wildlife monitoring.

Main Methods:

  • A novel NAS method was designed to operate directly on camera trap images with reduced resolutions and preserved aspect ratios.
  • The search process was guided by a custom loss function with a theoretically derived hyperparameter aimed at identifying lightweight network architectures.
  • The developed NAS method was applied to two distinct camera trap image datasets.

Main Results:

  • The NAS method successfully generated lightweight CNNs optimized for edge device deployment.
  • These networks were tested on an NVIDIA Jetson X2 edge device, demonstrating competitive classification accuracies.
  • The approach enables the creation of efficient models without requiring deep expertise in neural network design.

Conclusions:

  • The novel NAS method provides a practical and cost-effective solution for optimizing AI models for wildlife camera trap applications.
  • Researchers can leverage this method to deploy or expand wildlife surveillance systems on edge devices efficiently.
  • This facilitates broader and more accessible ecological research through automated wildlife monitoring.