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: Jul 4, 2025

Estimation of Contact Regions Between Hands and Objects During Human Multi-Digit Grasping
09:41

Estimation of Contact Regions Between Hands and Objects During Human Multi-Digit Grasping

Published on: April 21, 2023

1.6K

Improved 3D Markerless Mouse Pose Estimation Using Temporal Semi-Supervision.

Tianqing Li1, Kyle S Severson2, Fan Wang2

  • 1Duke University, Pratt School of Engineering, Department of Biomedical Engineering, Durham, 27708, NC, USA.

International Journal of Computer Vision
|January 26, 2024
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

[Developmental status and prospect of musical electroacupuncture].

Zhongguo zhen jiu = Chinese acupuncture & moxibustion·2015
Same author

Utility of Tc-PEG4-E[PEG4-c(RGDfK)]2 in Posttherapy Surveillance of Patients with Reelevated Carcinoembryonic Antigen Levels.

Medical principles and practice : international journal of the Kuwait University, Health Science Centre·2015
Same author

Characterization of the impurities and isomers in cefetamet pivoxil hydrochloride by liquid chromatography/time-of-flight mass spectrometry and ion trap mass spectrometry.

Journal of pharmaceutical and biomedical analysis·2015
Same author

(68)Ga-labeled 3PRGD2 for dual PET and Cerenkov luminescence imaging of orthotopic human glioblastoma.

Bioconjugate chemistry·2015
Same author

An exploratory study on 99mTc-RGD-BBN peptide scintimammography in the assessment of breast malignant lesions compared to 99mTc-3P4-RGD2.

PloS one·2015
Same author

Chemoradiation therapy reduces aldehyde dehydrogenase 1 expression in cervical cancer but does not improve patient survival.

Medical oncology (Northwood, London, England)·2015
Same journal

A Guide to Structureless Visual Localization.

International journal of computer vision·2026
Same journal

Distillation-free Scaling of Large State-Space Models for Images and Videos.

International journal of computer vision·2026
Same journal

Are Minimal Radial Distortion Solvers Really Necessary for Relative Pose Estimation?

International journal of computer vision·2026
Same journal

Structure-from-motion in micro-image domain for uncalibrated plenoptic 2.0 cameras.

International journal of computer vision·2026
Same journal

FourierMIL: Fourier Filtering-based Multiple Instance Learning for Whole Slide Image Analysis.

International journal of computer vision·2025
Same journal

A Likelihood Ratio-Based Approach to Segmenting Unknown Objects.

International journal of computer vision·2025
See all related articles

This study introduces a semi-supervised learning method for 3D animal pose estimation, improving accuracy and stability in tracking freely moving animals using multi-view video. The technique leverages unlabeled video data to enhance performance beyond current state-of-the-art methods.

Area of Science:

  • Animal behavior quantification
  • Computer vision
  • Machine learning for biological research

Background:

  • Markerless three-dimensional (3D) pose estimation from multi-view video is a promising technique for quantifying animal behavior.
  • Current methods face challenges due to limited training data and algorithms not optimized for animal-specific body plans.
  • Fully supervised convolutional neural networks (CNNs) require extensive labeled datasets for accurate 3D tracking.

Purpose of the Study:

  • To develop a more efficient and accurate method for 3D animal pose estimation.
  • To overcome the limitations of data scarcity in supervised learning approaches.
  • To improve the temporal stability and skeletal consistency of animal tracking.

Main Methods:

  • A semi-supervised learning strategy was developed, incorporating unlabeled video frames into the training process.

More Related Videos

Author Spotlight: Deciphering Electrical Networks Behind Complex Brain Activities and Disorders
05:49

Author Spotlight: Deciphering Electrical Networks Behind Complex Brain Activities and Disorders

Published on: November 1, 2024

791
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

10.7K

Related Experiment Videos

Last Updated: Jul 4, 2025

Estimation of Contact Regions Between Hands and Objects During Human Multi-Digit Grasping
09:41

Estimation of Contact Regions Between Hands and Objects During Human Multi-Digit Grasping

Published on: April 21, 2023

1.6K
Author Spotlight: Deciphering Electrical Networks Behind Complex Brain Activities and Disorders
05:49

Author Spotlight: Deciphering Electrical Networks Behind Complex Brain Activities and Disorders

Published on: November 1, 2024

791
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

10.7K
  • A simple temporal constraint was applied during training to leverage unlabeled data.
  • The method was evaluated on freely moving mice using multi-view video recordings.

    • Main Results:

    • The proposed semi-supervised approach significantly improved the state-of-the-art performance in multi-view volumetric 3D pose estimation.
    • Enhanced temporal stability and skeletal consistency were observed in the 3D tracking results.
    • The method demonstrated effectiveness in quantifying the behavior of freely moving animals.

    Conclusions:

    • Semi-supervised learning offers a viable solution to improve 3D animal pose estimation with reduced reliance on labeled data.
    • The developed technique advances the accuracy and reliability of markerless animal tracking systems.
    • This approach has the potential to accelerate behavioral research by providing more robust pose estimation tools.