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: Mar 24, 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

9.7K

Semi-Supervised Image-to-Video Adaptation for Video Action Recognition.

Jianguang Zhang, Yahong Han, Jinhui Tang

    IEEE Transactions on Cybernetics
    |March 19, 2016
    PubMed
    Summary

    This study enhances video action recognition by transferring knowledge from images, addressing limited video data. The novel approach improves accuracy and reduces overfitting, outperforming existing methods.

    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

    Reply to 'Immortal time bias limits the conclusion that post-transplant hypomethylating agent maintenance improves outcomes in myelodysplastic syndromes'.

    British journal of haematology·2026
    Same author

    Extracellular vesicle-mediated GALNT1/RPRD1A glycosylation axis drives immune escape and peritoneal metastasis in gastric cancer.

    Journal of experimental & clinical cancer research : CR·2026
    Same author

    Community-based chest X-ray screening versus symptom-based case presentation in the elderly: evidence from three consecutive years of data.

    BMC geriatrics·2026
    Same author

    MonSter++: Unified Stereo Matching, Multi-View Stereo, and Real-Time Stereo With Monodepth Priors.

    IEEE transactions on pattern analysis and machine intelligence·2026
    Same author

    Clinical characteristics and antibody responses to Omicron variants among pregnant women in China during the December 2022-April 2023 COVID-19 pandemic wave.

    Frontiers in immunology·2026
    Same author

    Spatial-Temporal Self-Compensating Graph Convolutional Network for Skeleton-Based Action Recognition Under Data Constraints.

    IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

    Area of Science:

    • Computer Vision
    • Machine Learning

    Background:

    • Human action recognition is crucial in computer vision.
    • Existing methods often lack sufficient labeled video data, leading to overfitting.
    • Knowledge transfer between different media types (images and videos) for action recognition is underexplored.

    Purpose of the Study:

    • To enhance video action recognition performance by adapting knowledge from images.
    • To address the challenge of limited labeled training videos in video action recognition.
    • To develop a semi-supervised framework leveraging both labeled and unlabeled videos.

    Main Methods:

    • Proposing an adaptation method to transfer action knowledge from images to videos.
    • Exploring common components between labeled videos and images to learn correlated action semantics.

    Related Experiment Videos

    Last Updated: Mar 24, 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

    9.7K
  • Extending the adaptation method into a semi-supervised framework.

    • Main Results:

    • The proposed method effectively adapts knowledge from images to improve video action recognition.
    • The semi-supervised framework alleviates overfitting by utilizing unlabeled videos.
    • Experiments demonstrate superior performance compared to state-of-the-art methods on benchmark and real-world datasets.

    Conclusions:

    • Knowledge adaptation from images significantly boosts video action recognition.
    • Semi-supervised learning combined with knowledge transfer is effective for handling limited labeled data.
    • The developed method offers a promising solution for robust human action recognition in videos.