Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Apr 12, 2026

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Published on: December 15, 2023

Unsupervised industrial image defect detection based on autoencoder and GANs.

Shuangli An¹, Junjie Wu¹, Jiawang Li¹

¹School of Intelligent Manufacturing and Control Engineering, Shanghai Polytechnic University, Shanghai, China.

|April 10, 2026

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

Neuropilin‑1: Emerging roles in nerve‑vessel‑bone coupling during fracture repair (Review).

International journal of molecular medicine·2026

Same author

Epimedii Folium Supplementation Improves Semen Quality, Hormonal Profile, and Immune Function by Modulating Gut Microbiota and Seminal Metabolites in Aged Boars.

Animals : an open access journal from MDPI·2026

Same author

Comprehensive evaluation of a deep learning-based synthetic CT model for MR-only radiotherapy across multiple anatomical sites.

Physics and imaging in radiation oncology·2026

Same author

Genetic Analysis of Pitt-Hopkins Syndrome Caused by a Novel Splicing Variant (c.1146+3A>T) in the TCF4 Gene.

Molecular genetics & genomic medicine·2026

Same author

Transcranial Magnetic Stimulation over the Left Inferior Parietal Lobule Facilitates Early-Stage Processing During Natural Chinese-English Bilingual Reading.

Brain sciences·2026

Same author

Effects of <i>Epimedium</i> Ultrafine Powder on Seminal Quality, Hormones, Immuno-Antioxidant Status, Gut Microbiota and Metabolites in Boars.

Animals : an open access journal from MDPI·2026

Same journal

Analysis of strength degradation of coal and rock masses and stability of mined areas under long term immersion environment.

PloS one·2026

Same journal

Biogenic Silver-Selenium nanocomposite with anticancer activity and potent efficacy against vancomycin-resistant Staphylococcus aureus.

PloS one·2026

Same journal

Preparation and physicochemical characterization of a biodegradable chitosan/carboxymethyl cellulose hydrogel synthesized in NaOH/urea medium.

PloS one·2026

Same journal

Action-guilt, survivor-guilt, and depression in combat-related PTSD.

PloS one·2026

Same journal

Explainable machine learning for predicting activities of daily living at discharge in stroke patients: A retrospective study using SHAP interpretability.

PloS one·2026

Same journal

Deep learning based two-way feature depiction model for brain tumor detection.

PloS one·2026

See all related articles

This study introduces an unsupervised industrial image defect detection method using autoencoders and Generative Adversarial Networks (GANs) to address limitations in traditional quality control. The novel approach enhances detection accuracy and generalization for manufacturing defect identification.

Area of Science:

Manufacturing Process Control
Artificial Intelligence in Industry
Computer Vision for Quality Assurance

Background:

Traditional defect detection methods in manufacturing struggle with large labeled datasets, weak generalization, and high costs.
Intelligent manufacturing demands high-precision, efficient quality control, especially in automotive, semiconductor, and electronics assembly.
Existing methods are inadequate for small or zero-sample defect detection scenarios, hindering robustness and adaptability.

Purpose of the Study:

To develop an unsupervised industrial image defect detection method for small and zero-sample scenarios.
To improve detection sensitivity, localization accuracy, and generalization ability for unknown defects.
To overcome the limitations of traditional supervised learning methods in industrial quality control.

Related Experiment Videos

Last Updated: Apr 12, 2026

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Published on: December 15, 2023

Main Methods:

Proposed a Multi-level Deep feature Adaptive fusion AutoEncoder (MDAAE) module integrated with Generative Adversarial Networks (GANs).
Utilized Pre-Trained Convolutional Neural Backbone Networks (PTCNBN) for multi-scale feature extraction.
Incorporated an Attention Mechanism (AM) for dynamic feature weighting, feature fusion, reconstruction, and enhanced GAN training via self-attention.

Main Results:

Achieved an Area Under the Curve (AUC) of 93.6 ± 0.5% and an F1-Score exceeding 0.890 ± 0.003 for defects like scratches and dents.
Demonstrated stable inference time within 3.0 ± 0.2 GB CPU memory and a converged Fréchet Inception Distance (FID) of 3.
Reported a low false alarm rate of 8.6 ± 0.7% under strong light conditions, indicating high robustness.

Conclusions:

The proposed unsupervised industrial defect detection method (UIIDD) exhibits strong robustness, generalization, reliability, and efficiency.
Effectively addresses the challenges of poor robustness, weak generalization, and high costs associated with traditional defect detection techniques.
Offers a viable solution for intelligent manufacturing quality control, particularly in scenarios with limited labeled data.