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: Apr 24, 2026

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

1.3K

Reparameterization-Driven Depthwise Separable Large-Kernel Network for Lightweight Salient Object Detection of Strip

Xiaofei Zhou, Zhenkun Mo, Gongyang Li

    IEEE Transactions on Cybernetics
    |April 22, 2026
    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

    A systematic review of forward simulation-based methods for point-source water pollution localization.

    Journal of contaminant hydrology·2026
    Same author

    Mitochondrial dysfunction via oxidative stress and calcium imbalance mediates antifungal activity of calcium propionate against Trichothecium roseum.

    International journal of food microbiology·2026
    Same author

    Probabilistic-based Learning for Joint Light Field Image Compression and Enhancement under Low-Light Conditions.

    IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026
    Same author

    Cuminal-triggered mitophagy induces mitochondrial dysfunction, resulting in decreased vegetative growth and pathogenicity in Trichothecium roseum.

    International journal of food microbiology·2026
    Same author

    Outpatient step-up dosing of bispecific antibodies in relapsed or refractory multiple myeloma: an oncology nursing framework for monitoring and supportive care.

    Frontiers in oncology·2026
    Same author

    DCPM-ADMET: fusion of dual-component pre-trained model and molecular fingerprints to enhance drug ADMET properties prediction.

    Journal of cheminformatics·2026

    This study introduces RepDSLKNet, a lightweight neural network for strip steel surface defect detection. It achieves high accuracy and efficiency, outperforming existing methods with significantly reduced computational cost.

    Area of Science:

    • Computer Vision
    • Deep Learning
    • Materials Science

    Background:

    • Strip steel surface defect detection is crucial for quality control.
    • Current deep learning models face a trade-off between accuracy and efficiency.
    • High-performing models are computationally expensive, while lightweight models lack accuracy.

    Purpose of the Study:

    • To develop a novel lightweight neural network for efficient and accurate strip steel surface defect detection.
    • To address the accuracy-efficiency trade-off in existing computer vision models.
    • To introduce a new model architecture that enhances feature extraction and fusion.

    Main Methods:

    • Proposed a spatial channel enhancement (SCE) module with reparameterizable depthwise large-kernel and pointwise convolutions.

    Related Experiment Videos

    Last Updated: Apr 24, 2026

    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

    1.3K
  • Developed a lightweight saliency model, RepDSLKNet, using SCE modules for encoder-decoder structure.
  • Integrated cascaded channel attention (CCA) modules for effective feature fusion.

    • Main Results:

    • RepDSLKNet achieves high detection accuracy with a small model size (0.47M parameters, 0.42G FLOPs).
    • Demonstrated a 19-fold improvement in throughput and a twofold reduction in latency compared to state-of-the-art methods.
    • Achieved competitive performance on public strip steel defect datasets.

    Conclusions:

    • RepDSLKNet offers an effective solution for strip steel surface defect detection, balancing accuracy and efficiency.
    • The proposed SCE module enhances feature extraction and strengthens spatial-channel interactions.
    • The lightweight architecture is suitable for real-world industrial applications requiring fast and accurate defect identification.