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: May 26, 2026

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

Geometry-aware localization evaluation of grad-CAM for wafer map defect classification.

Tushar Dudeja1, Prachi Sharma2

  • 1School of Electronics Engineering, Vellore Institute of Technology, Vellore, Tamil Nadu, India.

Scientific Reports
|May 24, 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

Association of hyponatremia and risk of in-hospital mortality in patients with acute stroke: a systematic review and meta-analysis.

Frontiers in stroke·2026
Same author

Clinical predictors of intubation and prolonged ventilation in pediatric Guillain-Barré syndrome requiring intensive care: A retrospective observational study.

Brain & development·2026
Same author

Frailty-driven anesthesia decisions in recurrent chronic subdural hematoma: Is it time to embrace embolization under local anesthesia?

Journal of anaesthesiology, clinical pharmacology·2026
Same author

Comparative evaluation of marginal bone loss and prosthetic success in immediate versus delayed implant loading protocols: A radiographic study.

Bioinformation·2026
Same author

2- and 4-quinolones as emerging anticancer scaffolds: Recent synthetic developments, SAR insights, and mechanistic perspective.

European journal of medicinal chemistry·2026
Same author

Prognostic Utility of Noninvasive Brain Monitoring in Moderate-to-Severe Cerebral Venous Thrombosis: A Prospective Observational Study.

Journal of neurosurgical anesthesiology·2026
Same journal

Therapeutic potential of crude protein extracts from two Egyptian freshwater snails Lanistes carinatus and Bellamya unicolor.

Scientific reports·2026
Same journal

Microbial contamination of donor corneas and post-keratoplasty endophthalmitis: a comparison between Japanese and U.S. eye banks using cold storage.

Scientific reports·2026
Same journal

Prevalence and contributing factors of virological non-suppression among adult patients on first-line antiretroviral therapy in tertiary hospitals in Ethiopia.

Scientific reports·2026
Same journal

An in vitro comparison of color stability between alkasite and different restorative materials in various staining solutions.

Scientific reports·2026
Same journal

Toward accessible mRNA LNP formulation: systematic evaluation of mixing strategies and key parameters.

Scientific reports·2026
Same journal

A network analysis of personality traits, mentalizing, and psychological health in Chinese college students.

Scientific reports·2026
See all related articles

Explainable AI (XAI) localization accuracy for semiconductor wafer inspection was improved by addressing radial attribution bias. A novel radial suppression method enhanced defect localization performance, particularly for edge-ring defects.

Area of Science:

  • Artificial Intelligence
  • Computer Vision
  • Semiconductor Manufacturing

Background:

  • Explainable AI (XAI) techniques like Grad-CAM are crucial for understanding model decisions in industrial inspection.
  • Evaluating XAI localization accuracy on large-scale, real-world datasets like WM811K is essential for practical application.
  • Existing XAI methods may exhibit biases that hinder precise defect localization in semiconductor wafer inspection.

Purpose of the Study:

  • To assess the localization accuracy of Grad-CAM and XGrad-CAM for semiconductor wafer defect detection.
  • To identify and characterize attribution biases in XAI methods applied to wafer defect datasets.
  • To develop and evaluate a method for mitigating radial attribution bias to improve localization performance.

Main Methods:

More Related Videos

Medical-grade Sterilizable Target for Fluid-immersed Fetoscope Optical Distortion Calibration
07:03

Medical-grade Sterilizable Target for Fluid-immersed Fetoscope Optical Distortion Calibration

Published on: February 23, 2017

Related Experiment Videos

Last Updated: May 26, 2026

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

Medical-grade Sterilizable Target for Fluid-immersed Fetoscope Optical Distortion Calibration
07:03

Medical-grade Sterilizable Target for Fluid-immersed Fetoscope Optical Distortion Calibration

Published on: February 23, 2017

  • Utilized the WM811K semiconductor wafer defect dataset (25,519 labeled maps).
  • Measured localization accuracy using Top-10% Intersection over Union (IoU) on a held-out test set.
  • Implemented and tested a radial suppression method to counteract observed radial attribution bias.

Main Results:

  • Baseline Top-10% IoU accuracy was 0.129, with consistent radial attribution bias observed in Grad-CAM and XGrad-CAM.
  • The radial suppression method improved average Top-10% IoU by 19.4% (to 0.155), statistically verified.
  • Significant localization improvement was noted for edge-ring defects, while other defect types showed stable performance.

Conclusions:

  • Grad-CAM and XGrad-CAM exhibit radial attribution bias in wafer defect localization, likely stemming from model representations.
  • A radial suppression technique effectively mitigates this bias, enhancing overall localization accuracy.
  • Geometric characteristics significantly influence residual localization errors, indicating areas for future refinement.