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: Sep 22, 2025

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
06:37

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

4.2K

Structured Domain Adaptation With Online Relation Regularization for Unsupervised Person Re-ID.

Yixiao Ge, Feng Zhu, Dapeng Chen

    IEEE Transactions on Neural Networks and Learning Systems
    |May 18, 2022
    PubMed
    Summary

    This study introduces a novel framework for unsupervised domain adaptation in open-set person re-identification. By incorporating relation-consistency regularization into domain translation, the method significantly enhances model performance on cross-domain tasks.

    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

    The signaling and regulatory functions of dying cells and cell corpses.

    Communications biology·2026
    Same author

    Dynamin-Related Protein 1 (Drp1) in Inflammatory Bowel Disease: Molecular Pathways Connecting Mitochondrial Dynamics with Intestinal Inflammation and Homeostasis.

    International journal of molecular sciences·2026
    Same author

    Electrophysiological and morphological alteration in the visual pathway of children with attention-deficit/hyperactivity disorder.

    Frontiers in psychiatry·2026
    Same author

    Risk factors for BPD and relevant nursing models of care to improve outcomes: a narrative review.

    Frontiers in pediatrics·2026
    Same author

    LGMN<sup>+</sup> macrophage promotes the formation of a tumor-supportive microenvironment in gastric cancer.

    Frontiers in immunology·2026
    Same author

    Therapeutic Targeting of Oxidative Phosphorylation in Microsatellite Instability-High Gastric Cancer.

    Journal of Cancer·2026

    Area of Science:

    • Computer Vision
    • Machine Learning
    • Artificial Intelligence

    Background:

    • Unsupervised Domain Adaptation (UDA) addresses model performance gaps between labeled source and unlabeled target domains.
    • Open-set person re-identification (re-ID) presents unique challenges due to non-overlapping identities across domains.
    • Existing domain translation methods for UDA in re-ID often underperform due to a lack of proper regularization.

    Purpose of the Study:

    • To propose a novel end-to-end structured domain adaptation framework for open-set person re-ID.
    • To introduce an online relation-consistency regularization term to improve domain translation.
    • To enhance feature encoder performance by leveraging both translated images and pseudo labels.

    Main Methods:

    • Developed an end-to-end framework integrating domain translation with relation-consistency regularization.

    Related Experiment Videos

    Last Updated: Sep 22, 2025

    Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
    06:37

    Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

    Published on: December 15, 2023

    4.2K
  • Optimized a person feature encoder to model inter-sample relations for supervising domain translation.
  • Employed pseudo-labeling, combining source-to-target translated images and target-domain images for joint training.

    • Main Results:

    • Achieved state-of-the-art performance on multiple unsupervised domain adaptation person re-ID benchmarks.
    • Demonstrated the effectiveness of relation-consistency regularization in domain translation for re-ID.
    • Secured second place in the Visual Domain Adaptation Challenge (VisDA) 2020 using synthetic-to-real translated images.

    Conclusions:

    • The proposed structured domain adaptation framework significantly advances open-set person re-ID.
    • Relation-consistency regularization is crucial for effective domain translation in UDA.
    • The framework offers a promising direction for improving cross-domain generalization in computer vision tasks.