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

A Seismic Data Acquisition System Based on Wireless Network Transmission.

Sensors (Basel, Switzerland)·2021
See all related articles

Related Experiment Video

Updated: Oct 16, 2025

Comprehensive Characterization of Extended Defects in Semiconductor Materials by a Scanning Electron Microscope
11:14

Comprehensive Characterization of Extended Defects in Semiconductor Materials by a Scanning Electron Microscope

Published on: May 28, 2016

14.0K

Cost-Sensitive Siamese Network for PCB Defect Classification.

Yilin Miao1, Zhewei Liu1, Xiangning Wu1,2

  • 1School of Computer Science, China University of Geosciences, Wuhan 430078, China.

Computational Intelligence and Neuroscience
|October 22, 2021
PubMed
Summary

This study introduces a cost-sensitive siamese network (CSS-Net) to accurately differentiate true and pseudo defects in printed circuit board (PCB) manufacturing. The novel model significantly improves defect detection accuracy while reducing inspection time and costs.

More Related Videos

Signal Acquisition, Score Interpretation, and Economics of a Non-Invasive Point-of-Care Test for Coronary Artery Disease
06:16

Signal Acquisition, Score Interpretation, and Economics of a Non-Invasive Point-of-Care Test for Coronary Artery Disease

Published on: August 9, 2024

580
Fully Automated Centrifugal Microfluidic Device for Ultrasensitive Protein Detection from Whole Blood
08:58

Fully Automated Centrifugal Microfluidic Device for Ultrasensitive Protein Detection from Whole Blood

Published on: April 16, 2016

10.7K

Related Experiment Videos

Last Updated: Oct 16, 2025

Comprehensive Characterization of Extended Defects in Semiconductor Materials by a Scanning Electron Microscope
11:14

Comprehensive Characterization of Extended Defects in Semiconductor Materials by a Scanning Electron Microscope

Published on: May 28, 2016

14.0K
Signal Acquisition, Score Interpretation, and Economics of a Non-Invasive Point-of-Care Test for Coronary Artery Disease
06:16

Signal Acquisition, Score Interpretation, and Economics of a Non-Invasive Point-of-Care Test for Coronary Artery Disease

Published on: August 9, 2024

580
Fully Automated Centrifugal Microfluidic Device for Ultrasensitive Protein Detection from Whole Blood
08:58

Fully Automated Centrifugal Microfluidic Device for Ultrasensitive Protein Detection from Whole Blood

Published on: April 16, 2016

10.7K

Area of Science:

  • Computer Science
  • Electrical Engineering
  • Manufacturing Technology

Background:

  • Printed circuit board (PCB) manufacturing requires rigorous defect detection.
  • Manual inspection is labor-intensive and time-consuming.
  • Automatic optical inspection (AOI) suffers from high false alarm rates, necessitating human intervention.

Purpose of the Study:

  • To develop a cost-sensitive deep learning model for distinguishing true and pseudo PCB defects.
  • To address the limitations of traditional AOI methods in PCB quality control.

Main Methods:

  • Proposed a cost-sensitive siamese network (CSS-Net) integrating siamese network, transfer learning, and threshold moving.
  • Framed PCB defect classification as a cost-sensitive problem.
  • Utilized optimization algorithms like NSGA-II to determine optimal cost-sensitive thresholds.

Main Results:

  • Achieved 97.60% accuracy in predicting true PCB defects.
  • Maintained 61.24% accuracy in predicting pseudo defects in a real-world production setting.
  • Outperformed state-of-the-art competitor models in comprehensive cost-sensitive metrics.
  • Reduced training time by an average of 33.32%.

Conclusions:

  • CSS-Net effectively distinguishes between true and pseudo PCB defects, enhancing manufacturing quality control.
  • The model offers a more efficient and accurate alternative to traditional inspection methods.
  • This approach significantly improves the cost-effectiveness of PCB defect detection.