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 10, 2026

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments
03:31

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments

Published on: December 15, 2023

Structure-Guided Domain-Adaptive Network for Few-Shot SAR Ship Detection.

Kang Ni, Weihang Zhou

    IEEE Transactions on Neural Networks and Learning Systems
    |June 8, 2026
    PubMed
    Summary

    This study introduces a novel Structure-Guided Domain-Adaptive Network (SGDANet) for few-shot synthetic aperture radar (SAR) ship detection. SGDANet effectively utilizes optical image features to enhance SAR ship detection accuracy, even with limited data.

    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

    Spatial patterns and key drivers of available phosphorus in Yunnan tea plantation soils.

    Journal of environmental management·2026
    Same author

    Camrelizumab Combined with Gemcitabine and Albumin-Bound Paclitaxel in Pancreatic Cancer Patients with Liver Metastases: A Prospective, Pilot Trial.

    Drug design, development and therapy·2026
    Same author

    Tea Cultivar Genotype Shapes Rhizosphere Microbiome Assembly Through Metabolic Differentiation.

    Plants (Basel, Switzerland)·2026
    Same author

    Endothelial Cell-Based Vascular Bandages for Blood-Brain Barrier Repair and Targeted siRNA Delivery.

    Advanced science (Weinheim, Baden-Wurttemberg, Germany)·2026
    Same author

    Precisely targeting of engineered nanovesicles to implanted scaffolds via click chemistry for microenvironment regulation in acute spinal cord injury.

    Journal of controlled release : official journal of the Controlled Release Society·2025
    Same author

    An engineering-reinforced extracellular vesicle-integrated hydrogel with an ROS-responsive release pattern mitigates spinal cord injury.

    Science advances·2025

    Area of Science:

    • Remote Sensing
    • Computer Vision
    • Artificial Intelligence

    Background:

    • Synthetic Aperture Radar (SAR) ship detection is challenged by complex backgrounds and limited annotated samples.
    • Optical remote sensing images offer high resolution and intuitive visualization, complementing SAR data.
    • Existing methods struggle with feature representation and domain adaptation for SAR ship detection.

    Purpose of the Study:

    • To propose a novel network, SGDANet, for few-shot SAR ship detection.
    • To leverage structural features from optical images to guide SAR target feature learning.
    • To improve domain adaptation and generalization capabilities in SAR ship detection.

    Main Methods:

    • Developed a convolutional neural networks (CNNs)-transformer architecture (SGDANet).

    Related Experiment Videos

    Last Updated: Jun 10, 2026

    End-To-End Deep Neural Network for Salient Object Detection in Complex Environments
    03:31

    End-To-End Deep Neural Network for Salient Object Detection in Complex Environments

    Published on: December 15, 2023

  • Modeled and embedded structural and edge token features from optical images.
  • Implemented a split-fuse-merge strategy with an attention mechanism for feature fusion and adversarial domain adaptation.

    • Main Results:

    • SGDANet significantly outperformed existing models in three-shot and five-shot SAR ship detection scenarios.
    • The network demonstrated strong performance in zero-shot SAR target detection, indicating excellent generalization.
    • Experiments were conducted on three self-built ship datasets.

    Conclusions:

    • SGDANet effectively addresses challenges in few-shot SAR ship detection by integrating optical image structural guidance.
    • The proposed feature fusion and domain adaptation mechanisms enhance detection accuracy and robustness.
    • SGDANet shows significant potential for real-world applications requiring reliable SAR ship detection with limited data.