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: May 14, 2025

3D Kinematic Gait Analysis for Preclinical Studies in Rodents
10:19

3D Kinematic Gait Analysis for Preclinical Studies in Rodents

Published on: August 3, 2019

10.6K

GCNTrack: A Pig-Tracking Method Based on Skeleton Feature Similarity.

Zhaoyang Yin1, Zehua Wang1, Junhua Ye2

  • 1School of Mathematics and Computer Science, Zhejiang Agriculture and Forestry University, Hangzhou 311300, China.

Animals : an Open Access Journal From MDPI
|April 12, 2025
PubMed
Summary
This summary is machine-generated.

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

An Automatic Movement Monitoring Method for Group-Housed Pigs.

Animals : an open access journal from MDPI·2024
Same author

Deep Learning-Based Human Activity Real-Time Recognition for Pedestrian Navigation.

Sensors (Basel, Switzerland)·2020
See all related articles

GcnTrack improves pig tracking accuracy using skeleton features and a dual-tracking strategy, even with incomplete camera views. This method enhances pig behavior and health monitoring on farms.

Area of Science:

  • Animal Science
  • Computer Vision
  • Agricultural Technology

Background:

  • Accurate pig tracking is crucial for monitoring animal behavior and health.
  • Real-farm pig tracking faces challenges due to incomplete camera fields of view (FOV), leading to frequent entry/exit events that reduce accuracy.
  • Existing methods struggle with continuous tracking when pigs leave and re-enter the camera's FOV.

Purpose of the Study:

  • To develop an efficient and accurate pig-tracking method robust to incomplete camera FOV.
  • To improve the reliability of pig tracking for enhanced animal welfare and farm management.
  • To address the limitations of current tracking systems in dynamic farm environments.

Main Methods:

  • Proposed GcnTrack method utilizing skeleton feature similarity for pig tracking.
Keywords:
graph convolutional networkpigre-identificationskeletontracking

More Related Videos

FIM Imaging and FIMtrack: Two New Tools Allowing High-throughput and Cost Effective Locomotion Analysis
10:02

FIM Imaging and FIMtrack: Two New Tools Allowing High-throughput and Cost Effective Locomotion Analysis

Published on: December 24, 2014

11.7K
Tracking Mouse Bone Marrow Monocytes In Vivo
12:08

Tracking Mouse Bone Marrow Monocytes In Vivo

Published on: February 27, 2015

9.5K

Related Experiment Videos

Last Updated: May 14, 2025

3D Kinematic Gait Analysis for Preclinical Studies in Rodents
10:19

3D Kinematic Gait Analysis for Preclinical Studies in Rodents

Published on: August 3, 2019

10.6K
FIM Imaging and FIMtrack: Two New Tools Allowing High-throughput and Cost Effective Locomotion Analysis
10:02

FIM Imaging and FIMtrack: Two New Tools Allowing High-throughput and Cost Effective Locomotion Analysis

Published on: December 24, 2014

11.7K
Tracking Mouse Bone Marrow Monocytes In Vivo
12:08

Tracking Mouse Bone Marrow Monocytes In Vivo

Published on: February 27, 2015

9.5K
  • Employed YOLOv7-Pose for extracting pig skeleton key points.
  • Implemented a dual-tracking strategy combining IOU matching and graph convolutional reidentification (Re-ID) for continuous tracking, including pigs re-entering the FOV.

    • Main Results:

    • GcnTrack achieved 84.98% Multiple Object Tracking Accuracy (MOTA) and 82.22% identification F1 Score (IDF1) on short-duration videos.
    • Demonstrated 74% tracking precision on medium-duration videos.
    • On long-duration videos, pigs entered the scene 15.29 times on average with 6.28 identity switches per pig, indicating robust performance despite frequent occlusions.

    Conclusions:

    • GcnTrack offers an accurate and reliable solution for pig tracking in challenging farm environments with incomplete camera FOV.
    • The method effectively handles pigs entering and exiting the camera's view, ensuring continuous monitoring.
    • This advancement supports better assessment of pig behavior and health through improved tracking technology.