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: Oct 11, 2025

Author Spotlight: An Efficient and Robust Software for Automated Fusion of Multiple Preclinical Imaging Modalities
07:13

Author Spotlight: An Efficient and Robust Software for Automated Fusion of Multiple Preclinical Imaging Modalities

Published on: October 27, 2023

1.4K

Pixel and Feature Transfer Fusion for Unsupervised Cross-Dataset Person Reidentification.

Yang Yang, Guan'an Wang, Prayag Tiwari

    IEEE Transactions on Neural Networks and Learning Systems
    |December 1, 2021
    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

    Application of Artificial Intelligence In Drug-target Interactions Prediction: A Review.

    npj biomedical innovations·2026
    Same author

    Neural Network Optimization Reimagined: Decoupled Techniques for Scratch and Fine-Tuning.

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

    Enhancing anticancer peptide discovery: A fusion-centric framework with conditional diffusion for prediction and generation.

    PLoS computational biology·2026
    Same author

    MARINE-Transformer: A General-purpose framework for multivariate ocean time series analysis.

    Neural networks : the official journal of the International Neural Network Society·2026
    Same author

    Adaptive Boosting LLMs for Text Classification.

    IEEE transactions on neural networks and learning systems·2026
    Same author

    TAPB: an interventional debiasing framework for alleviating target prior bias in drug-target interaction prediction.

    Nature communications·2025
    Same journal

    ASSR-Net: Anisotropic Structure-Aware and Spectrally Recalibrated Network for Hyperspectral Image Fusion.

    IEEE transactions on neural networks and learning systems·2026
    Same journal

    PIMPC-GNN: Physics-Informed Multiphase Consensus Learning for Enhancing Imbalanced Node Classification in Graph Neural Networks.

    IEEE transactions on neural networks and learning systems·2026
    Same journal

    Quantum Rényi α-Entropies for Graph Characterization.

    IEEE transactions on neural networks and learning systems·2026
    Same journal

    LANet: A Lightweight and Accurate Balanced Network Based on State Space Models for Real-Time Semantic Segmentation.

    IEEE transactions on neural networks and learning systems·2026
    Same journal

    MENDNet: Memory-Enhanced Dependency Network for Multistock Movement Prediction.

    IEEE transactions on neural networks and learning systems·2026
    Same journal

    Temporal Mask-Embedding Learning and Query-Refined Head Network for Visual Tracking.

    IEEE transactions on neural networks and learning systems·2026
    See all related articles

    This study introduces a novel recurrent autoencoder (RAE) framework for unsupervised cross-dataset person reidentification (Re-ID). The RAE unifies pixel and feature alignment methods, achieving state-of-the-art results on benchmark datasets.

    Area of Science:

    • Computer Vision
    • Artificial Intelligence
    • Machine Learning

    Background:

    • Unsupervised cross-dataset person reidentification (Re-ID) is crucial for transferring knowledge from labeled source domains to unlabeled target domains.
    • Existing frameworks primarily rely on either pixel-alignment (low-level knowledge) or feature-alignment (high-level knowledge).

    Purpose of the Study:

    • To propose a novel recurrent autoencoder (RAE) framework that unifies and enhances both pixel-alignment and feature-alignment methods for unsupervised cross-dataset person Re-ID.
    • To improve the performance of person Re-ID by effectively transferring both low-level and high-level knowledge across datasets.

    Main Methods:

    • A recurrent autoencoder (RAE) framework comprising three modules: feature-transfer (FT), pixel-transfer (PT), and fusion.

    Related Experiment Videos

    Last Updated: Oct 11, 2025

    Author Spotlight: An Efficient and Robust Software for Automated Fusion of Multiple Preclinical Imaging Modalities
    07:13

    Author Spotlight: An Efficient and Robust Software for Automated Fusion of Multiple Preclinical Imaging Modalities

    Published on: October 27, 2023

    1.4K
  • The FT module maps images to a shared feature space, reducing domain gap and enhancing identity discriminability.
  • The PT module reconstructs images from target features in the source style, propagating low-level knowledge. A fusion module combines both knowledge types using bilinear pooling.

    • Main Results:

    • The proposed RAE framework significantly outperforms existing pixel-alignment and feature-alignment methods.
    • New state-of-the-art results were achieved on the Market-1501, DukeMTMC-ReID, and MSMT17 person Re-ID datasets.

    Conclusions:

    • The unified RAE framework effectively transfers both high-level and low-level knowledge for unsupervised cross-dataset person Re-ID.
    • This approach offers a more robust and effective solution compared to methods relying on a single alignment strategy.