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 7, 2026

Trajectory Data Analyses for Pedestrian Space-time Activity Study
16:14

Trajectory Data Analyses for Pedestrian Space-time Activity Study

Published on: February 25, 2013

14.3K

Video-Based Pedestrian Re-Identification by Adaptive Spatio-Temporal Appearance Model.

Wei Zhang, Bingpeng Ma, Kan Liu

    IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society
    |February 27, 2017
    PubMed
    Summary

    This study introduces a novel spatiotemporal approach for pedestrian re-identification, improving accuracy by analyzing body-action units over time. The method effectively captures temporal variations in appearance for better person recognition in videos.

    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

    Overnight sleep features and next-morning brain metabolism in older adults.

    Sleep medicine·2026
    Same author

    SERPINE2-mediated activation of JAK2/STAT3 facilitates NRF2 nuclear translocation and GCLC transcription to confer ferroptosis resistance and lenvatinib resistance in hepatocellular carcinoma.

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

    Tumor microenvironment-driven mechanisms of photodynamic therapy resistance and emerging targeted combination strategies.

    Journal of photochemistry and photobiology. B, Biology·2026
    Same author

    Identification of a Prognostic Gene Signature Based on Lenvatinib Resistance in Hepatocellular Carcinoma with Functional Validation of the Key Gene CPB2.

    Journal of hepatocellular carcinoma·2026
    Same author

    Clinician engagement shapes the impact of AI-based ECG screening for chronic liver disease in primary care.

    NPJ digital medicine·2026
    Same author

    Artificial intelligence-enhanced electrocardiogram for the detection of rheumatic heart disease.

    European heart journal. Digital health·2026

    Area of Science:

    • Computer Vision
    • Artificial Intelligence
    • Pattern Recognition

    Background:

    • Pedestrian re-identification faces challenges due to appearance variations (pose, illumination, occlusion).
    • Existing spatial alignment methods treat body parts independently, overlooking temporal action phases.
    • Temporal dynamics of body parts during actions are crucial for accurate re-identification.

    Purpose of the Study:

    • To propose a novel spatiotemporal approach for pedestrian re-identification.
    • To address the limitations of spatial alignment by incorporating temporal information.
    • To develop a robust appearance representation for improved person recognition in videos.

    Main Methods:

    • Exploiting the periodicity of walking gaits to generate a spatiotemporal body-action model.

    More Related Videos

    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

    Related Experiment Videos

    Last Updated: Mar 7, 2026

    Trajectory Data Analyses for Pedestrian Space-time Activity Study
    16:14

    Trajectory Data Analyses for Pedestrian Space-time Activity Study

    Published on: February 25, 2013

    14.3K
    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
  • Decomposing the model into body-action units representing action primitives.
  • Learning and extracting Fisher vectors from these units for a comprehensive representation.
  • Globally aligning spatiotemporal appearance, unlike local dynamic features.

    • Main Results:

    • The proposed spatiotemporal representation significantly enhances pedestrian re-identification.
    • The approach effectively models appearance variations across different action phases.
    • Experimental results demonstrate superior performance compared to state-of-the-art methods on public datasets.

    Conclusions:

    • The integration of temporal alignment with spatial alignment offers a more effective solution for pedestrian re-identification.
    • The body-action unit model captures essential dynamic and appearance information.
    • This method provides a robust and globally aligned spatiotemporal feature representation for improved person recognition.