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 Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Critical slowing down of semiarid vegetation resilience is amplified by intensifying heatwaves.

Nature communications·2026
Same author

Measurement of Serum Parameters Attributes the Therapeutic Effects of Music Therapy to Augmented Stress-Coping Ability and Diminished Systemic Inflammation.

Depression and anxiety·2026
Same author

Why sycophantic LLMs may imperil interactive norms between humans.

Communications psychology·2026
Same author

Intermetallic Heusler Alloys for Catalysis.

Small methods·2026
Same author

Risk factors for early neurological deterioration in patients with acute ischaemic stroke and assessment of short-term prognosis.

Frontiers in neurology·2026
Same author

Electrochemical chlorination byproducts (ECBPs): A potential blind spot in electrochemical treatment of waste activated sludge.

Journal of hazardous materials·2026

Related Experiment Video

Updated: May 21, 2025

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

2.6K

LDMP-RENet: Reducing intra-class differences for metal surface defect few-shot semantic segmentation.

Jiyan Zhang1, Hanze Ding1, Zhangkai Wu2

  • 1College of Mathematics and Information Engineering, Longyan University, Longyan, China.

Plos One
|March 17, 2025
PubMed
Summary

Few-shot segmentation models struggle with metal surface defect variations. Our novel LDMP-RENet network effectively addresses semantic and distortion differences, improving defect detection accuracy in industrial settings.

More Related Videos

Scanning Electron Microscopic Evaluation of Surface Defects of Remover Retreatment File After Single and Multiple Uses
03:07

Scanning Electron Microscopic Evaluation of Surface Defects of Remover Retreatment File After Single and Multiple Uses

Published on: October 11, 2024

396
Evaporation-reducing Culture Condition Increases the Reproducibility of Multicellular Spheroid Formation in Microtiter Plates
11:24

Evaporation-reducing Culture Condition Increases the Reproducibility of Multicellular Spheroid Formation in Microtiter Plates

Published on: March 7, 2017

6.9K

Related Experiment Videos

Last Updated: May 21, 2025

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

2.6K
Scanning Electron Microscopic Evaluation of Surface Defects of Remover Retreatment File After Single and Multiple Uses
03:07

Scanning Electron Microscopic Evaluation of Surface Defects of Remover Retreatment File After Single and Multiple Uses

Published on: October 11, 2024

396
Evaporation-reducing Culture Condition Increases the Reproducibility of Multicellular Spheroid Formation in Microtiter Plates
11:24

Evaporation-reducing Culture Condition Increases the Reproducibility of Multicellular Spheroid Formation in Microtiter Plates

Published on: March 7, 2017

6.9K

Area of Science:

  • Computer Vision
  • Machine Learning
  • Materials Science

Background:

  • Few-shot segmentation models excel at data-insufficient tasks like metal defect detection.
  • Existing models overlook intra-class variations (semantic and distortion) in metal surface defects, limiting segmentation accuracy.
  • Accurate metal defect detection is crucial for industrial quality control.

Purpose of the Study:

  • To develop a novel network that addresses intra-class differences in metal surface defect segmentation.
  • To improve the generalization capability of few-shot segmentation models for industrial defect detection.
  • To achieve precise pixel-level segmentation of metal surface defects.

Main Methods:

  • Introduction of the Local Descriptor-based Multi-Prototype Reasoning and Excitation Network (LDMP-RENet).
  • Utilizing a multi-prototype reasoning module for local-view feature relevance (semantic difference obviation).
  • Employing a multi-prototype excitation module for global-view feature relevance (distortion difference obviation).
  • Integrating local and global views via an information fusion module for pixel-level mask generation.

Main Results:

  • The proposed LDMP-RENet network demonstrates superior performance compared to existing benchmarks on metal defect datasets.
  • The network effectively handles both semantic and distortion intra-class differences in metal surface defects.
  • Achieved state-of-the-art results in precise metal defect segmentation.

Conclusions:

  • LDMP-RENet significantly enhances few-shot segmentation for metal defect detection by addressing intra-class variations.
  • The two-view guidance (local and global information fusion) is key to precise segmentation.
  • This work sets a new benchmark for industrial defect detection using few-shot segmentation.