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: Nov 29, 2025

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

828

Zero-Shot Video Object Segmentation With Co-Attention Siamese Networks.

Xiankai Lu, Wenguan Wang, Jianbing Shen

    IEEE Transactions on Pattern Analysis and Machine Intelligence
    |November 24, 2020
    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

    Latent Chain-of-Thought for Visual Reasoning.

    Advances in neural information processing systems·2026
    Same author

    Enhanced Root Exudation as an Adaptation Mechanism to Facilitate Phosphorus Mobilization in a Primary Tropical Forest Under Chronic Nitrogen Deposition.

    Global change biology·2026
    Same author

    Mask-Guided Self-Supervised Video Object Segmentation.

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

    Recent Progress of Single-Ion Conducting Polymer Electrolytes for Rechargeable Mono- and Multivalent Cation-Based Metal Batteries.

    Angewandte Chemie (International ed. in English)·2026
    Same author

    REEs-HNMD weakens riverine N-sink function through suppressed sediment denitrification in a rare-earth mining watershed.

    Journal of hazardous materials·2026
    Same author

    Global hotspots of particulate organic carbon losses under climate change.

    Nature communications·2026

    We introduce COSNet, a novel network for zero-shot video object segmentation. This method leverages global co-attention to effectively capture frame correlations, significantly improving segmentation accuracy over existing approaches.

    Area of Science:

    • Computer Vision
    • Artificial Intelligence
    • Machine Learning

    Background:

    • Current deep learning methods for video object segmentation often focus on short-term temporal segments.
    • These methods may overlook inherent correlations among video frames, limiting performance.
    • A holistic approach is needed to capture global scene context and improve segmentation.

    Purpose of the Study:

    • To introduce a novel network, CO-attention Siamese Network (COSNet), for zero-shot video object segmentation.
    • To address limitations of existing methods by incorporating global co-attention mechanisms.
    • To provide a unified, end-to-end trainable framework for improved video object segmentation.

    Main Methods:

    • COSNet utilizes a global co-attention mechanism to exploit correlations among video frames.

    More Related Videos

    Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
    04:48

    Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

    Published on: November 30, 2022

    3.2K

    Related Experiment Videos

    Last Updated: Nov 29, 2025

    Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
    03:31

    Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

    Published on: December 15, 2023

    828
    Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
    04:48

    Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

    Published on: November 30, 2022

    3.2K
  • Co-attention layers capture global correlations and scene context within a joint feature space.
  • The network is trained with pairs or groups of video frames, augmenting training data and capacity.

    • Main Results:

    • COSNet demonstrates superior performance in zero-shot video object segmentation compared to state-of-the-art methods.
    • Experiments on three large benchmarks show significant improvements achieved by COSNet.
    • The co-attention model effectively infers salient foreground objects by processing multiple reference frames.

    Conclusions:

    • COSNet offers a novel and effective approach to zero-shot video object segmentation.
    • The global co-attention mechanism is key to capturing long-range dependencies and scene context.
    • COSNet represents a significant advancement in the field, outperforming existing alternatives.