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

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

Pointer Defect Detection Based on Transfer Learning and Improved Cascade-RCNN.

Sensors (Basel, Switzerland)·2020

Same author

Preparation and properties of copper-oil-based nanofluids.

Nanoscale research letters·2011

Same author

Ordered gelation of chemically converted graphene for next-generation electroconductive hydrogel films.

Angewandte Chemie (International ed. in English)·2011

Same author

A highly selective and sensitive on-off sensor for silver ions and cysteine by light scattering technique of DNA-functionalized gold nanoparticles.

Chemical communications (Cambridge, England)·2011

Same author

The remarkable enhancement of CO-pretreated CuO-Mn2O3/γ-Al2O3 supported catalyst for the reduction of NO with CO: the formation of surface synergetic oxygen vacancy.

Chemistry (Weinheim an der Bergstrasse, Germany)·2011

Same author

Application of multiwall carbon nanotubes-based matrix solid phase dispersion extraction for determination of hormones in butter by gas chromatography mass spectrometry.

Journal of chromatography. A·2011

Search research articles

Related Experiment Video

Updated: Nov 10, 2025

Author Spotlight: Efficient Image Recognition Using Directional Gradient Histogram Technique and Support Vector Machines

Author Spotlight: Efficient Image Recognition Using Directional Gradient Histogram Technique and Support Vector Machines

Published on: January 5, 2024

A New Steel Defect Detection Algorithm Based on Deep Learning.

Weidong Zhao¹, Feng Chen¹, Hancheng Huang¹

¹College of Electrical and Information Engineering, Anhui University of Technology, Maanshan 243000, China.

Computational Intelligence and Neuroscience

|April 7, 2021

Summary

This summary is machine-generated.

This study introduces an improved deep learning algorithm for detecting small and complex steel surface defects. The enhanced Faster R-CNN model achieves higher accuracy, improving steel quality control.

More Related Videos

Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography

Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography

Published on: May 15, 2017

Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model

Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model

Published on: August 16, 2020

Related Experiment Videos

Last Updated: Nov 10, 2025

Author Spotlight: Efficient Image Recognition Using Directional Gradient Histogram Technique and Support Vector Machines

Author Spotlight: Efficient Image Recognition Using Directional Gradient Histogram Technique and Support Vector Machines

Published on: January 5, 2024

Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography

Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography

Published on: May 15, 2017

Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model

Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model

Published on: August 16, 2020

Area of Science:

Materials Science
Computer Science
Artificial Intelligence

Background:

Deep learning advancements have spurred research in target detection algorithms.
Detecting small and complex targets, particularly steel surface defects, remains a significant challenge.
Existing algorithms often exhibit low detection accuracy on datasets like NEU-DET.

Purpose of the Study:

To address the limitations of current deep learning detection algorithms for small and complex steel surface defects.
To propose and validate an improved target detection algorithm for steel surface defect identification.
To enhance the accuracy and effectiveness of automated steel defect detection systems.

Main Methods:

Modified the traditional Faster R-CNN algorithm by reconstructing its network structure.
Implemented multi-scale fusion training to improve detection of small target features.
Integrated deformable convolution networks to better capture complex target features.

Main Results:

The proposed method achieved a mean average precision of 0.752, an improvement of 0.128 over the original algorithm.
Specific average precisions for defects include: crazing (0.501), inclusion (0.791), patches (0.792), pitted surface (0.874), rolled in scale (0.649), and scratches (0.905).
The enhanced model effectively identifies small and complex defects on steel surfaces.

Conclusions:

The developed deep learning network model demonstrates superior detection performance for steel surface defects.
The improved Faster R-CNN algorithm offers a viable solution for the automatic detection of steel defects.
This research provides a valuable reference for advancing automated quality control in steel production.