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 16, 2026

Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography
11:34

Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography

Published on: May 15, 2017

Lightweight metal surface defect detection algorithm based on pruning and knowledge distillation.

Yiqing Cao1, Yonger Yao2, Lijun Lu3

  • 1College of Intelligent Manufacturing, Putian University, Putian, 351100, Fujian, China. caoyiqing1987@163.com.

Scientific Reports
|June 14, 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

LLM-Enhanced Multimodal Fusion of SPECT Radiomics and Clinical Data for Predicting 131I Therapeutic Response in Differentiated Thyroid Cancer.

Molecular imaging and biology·2026
Same author

Does Warmed Saline Irrigation Improve Comfort During Dressing Changes After Digit Replantation? A Randomized Controlled Trial.

Advances in wound care·2026
Same author

P1-C<sub>7</sub>N<sub>4</sub>O<sub>8</sub>: improved stability and energy density <i>via</i> a covalent framework with uniform single bonds.

Physical chemistry chemical physics : PCCP·2026
Same author

Pan-cancer 8q24 amplification predicts primary immunotherapy resistance and therapeutic vulnerabilities.

iScience·2026
Same author

Cross-Tracer Synthesis Model of <sup>11</sup>C-CFT and <sup>18</sup>F-DOPA PET Images from <sup>18</sup>F-FDG for Parkinson's Disease.

IEEE journal of biomedical and health informatics·2026
Same author

Selective degradation mechanism induced by stacked structure of phenanthrene aggregates in thermally activated persulfate oxidation.

Journal of environmental sciences (China)·2026

This study introduces a lightweight algorithm for metal surface defect detection, significantly reducing model size and computational needs for industrial edge devices. The method achieves high accuracy while enabling real-time defect identification systems.

Area of Science:

  • Computer Vision
  • Machine Learning
  • Industrial Automation

Background:

  • Metal surface defect detection models often suffer from large parameters and high computational complexity.
  • This limits their deployment on resource-constrained industrial edge devices.

Purpose of the Study:

  • To develop a lightweight algorithm for efficient metal surface defect detection.
  • To enable the deployment of defect detection models on industrial edge devices.
  • To create a real-time interactive system for industrial defect detection.

Main Methods:

  • Proposed a lightweight algorithm based on YOLOv8, incorporating L1 regularization structured pruning for model compression.
  • Employed a collaborative pruning and distillation approach, using pruned (student) and unpruned (teacher) models.
Keywords:
Industrial inspectionKnowledge distillationMetal surface defect detectionModel pruningYOLOv8

Related Experiment Videos

Last Updated: Jun 16, 2026

Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography
11:34

Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography

Published on: May 15, 2017

  • Utilized a novel BFCD-KD knowledge distillation method to fuse middle-layer features with detection head branches and original model loss.

    • Main Results:

    • Reduced model FLOPs by 51.0% and parameters by 41.5%.
    • Achieved a mean Average Precision (mAP@0.50) of 70.6% on the GC10-DET dataset.
    • Developed a real-time interactive metal surface defect detection system for industrial sites.

    Conclusions:

    • The proposed lightweight algorithm effectively balances accuracy and efficiency for metal surface defect detection.
    • The developed system facilitates real-time visualization and interactive defect detection in industrial settings.
    • The approach enables the deployment of advanced AI models on edge devices for industrial automation.