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: Jun 23, 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

Multi-granularity dynamic hierarchical graphs for video-based person re-identification.

Wei Zhao1, Bingyi Zhou1, Yongquan Wang1

  • 1School of Artificial Intelligence and Automation, Huazhong University of Science and Technology, 1037 Luoyu Road, 430074, Wuhan, China.

Neural Networks : the Official Journal of the International Neural Network Society
|June 20, 2026
PubMed
Summary

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

From Subjective to Objective: The Bedside Value of HRV in AssessingChronic Pain Complicated by Emotionaland Sleep Issues.

Applied psychophysiology and biofeedback·2026
Same author

Next-generation strategies for anterior cruciate ligament repair: constructing biointelligent ligament grafts integrating biomimetic design, immune modulation, and sensory feedback.

Frontiers in bioengineering and biotechnology·2026
Same author

Gelatin macromolecular microspheres constructed by microfluidics regulate the TNF/HIF-1 signaling axis to reshape the bone immune microenvironment and enhance infectious bone defect repair.

Materials today. Bio·2026
Same author

Correspondence on 'Development of bispecific antibodies that drive selective targeted local complement inhibition on rheumatoid arthritis-related antigens' by Wang et al.

Annals of the rheumatic diseases·2026
Same author

The Gut-Bone Axis: Molecular Mechanisms and Therapeutic Perspectives in Skeletal Homeostasis.

Frontiers in bioscience (Landmark edition)·2026
Same author

Beyond muscle mass: methodological rigor and clinical translation in interventions for sarcopenic obesity with type 2 diabetes.

Nutrition & metabolism·2026

This study introduces a dynamic hierarchical graph network (DHGN) for video-based person re-identification (Re-ID). The novel approach effectively models temporal features, improving pedestrian identification accuracy even with occlusions.

Area of Science:

  • Computer Vision
  • Artificial Intelligence
  • Machine Learning

Background:

  • Video-based person re-identification (Re-ID) is crucial for surveillance and security systems.
  • Existing Re-ID methods struggle with occlusions and accurately modeling pedestrian motion due to fixed-length temporal feature extraction.

Purpose of the Study:

  • To propose a novel graph-based framework, the dynamic hierarchical graph network (DHGN), for improved video-based person Re-ID.
  • To address limitations in modeling temporal features and handling occlusions in existing Re-ID methods.

Main Methods:

  • Developed a dynamic hierarchical graph network (DHGN) that adaptively segments video features and constructs graphs based on feature similarity.
  • Employed graph nodes representing frame features at the same level, connected adaptively to capture multi-granularity temporal clues.
Keywords:
Feature inferenceGraph neural networkVideo-based person re-identification

Related Experiment Videos

Last Updated: Jun 23, 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

  • Introduced a similarity weighted inference module (SWIM) to enhance matching robustness by utilizing gallery-gallery similarity.

    • Main Results:

    • The DHGN framework adaptively captures temporal cues from different body parts by aggregating features from neighboring nodes.
    • The SWIM module refines query-gallery similarity, leading to more robust matching results.
    • Extensive experiments on four benchmarks demonstrated the significant effectiveness of the proposed DHGN method.

    Conclusions:

    • The proposed DHGN framework offers a superior approach to modeling temporal features for video-based person Re-ID.
    • The method effectively handles occlusions and improves the accuracy of identifying pedestrians in distributed camera systems.
    • The DHGN and SWIM modules collectively enhance the robustness and performance of person Re-ID systems.