Jove
Visualize
联系我们
JoVE
x logofacebook logolinkedin logoyoutube logo
关于 JoVE
概览领导团队博客JoVE 帮助中心
作者
出版流程编辑委员会范围与政策同行评审常见问题投稿
图书馆员
用户评价订阅访问资源图书馆顾问委员会常见问题
研究
JoVE JournalMethods CollectionsJoVE Encyclopedia of Experiments存档
教育
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab Manual教师资源中心教师网站
使用条款与条件
隐私政策
政策

相关概念视频

您也可能阅读

相关文章

通过共同作者、期刊和引用图与本文相关的文章。

排序
Same author

The effect of pectin-supplemented enteral nutrition in mechanically ventilated patients receiving gastric feeding: protocol of a multicenter, open-label, randomized controlled trial (the PROMOTE trial).

Frontiers in medicine·2026
Same author

GPRASP1 Safeguards Endothelial Aspartate Metabolism to Prevent Pulmonary Hypertension Associated With Heart Failure With Preserved Ejection Fraction.

Circulation·2026
Same author

Separation and Characterization of Self-Assembled Nanoparticles from <i>Rheum palmatum</i> L.-<i>Salvia miltiorrhiza</i> Bunge Extract and Their Renoprotective Effects in Acute Kidney Injury.

Antioxidants (Basel, Switzerland)·2026
Same author

BRD4 directs myofiber identity and metabolic adaptation through CHD4 cooperation.

Nature communications·2026
Same author

How social support impacts depression in Chinese emerging adults: mediators of cognitive reappraisal and loneliness.

Frontiers in psychology·2026
Same author

Soil microbial community assembly drives ecosystem multifunctionality under grazing disturbance by regulating diversity and network structure.

Scientific reports·2026

相关实验视频

Updated: Jun 12, 2025

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

485

一个轻量级的平行注意力残留网络,用于缺陷识别.

Cheng Lv1, Enxu Zhang1, Guowei Qi1

  • 1School of Mechanical Engineering, Xijing University, Xi'an, 710123, China.

Scientific reports
|September 19, 2024
PubMed
概括
此摘要是机器生成的。

这项研究引入了一种用于检测磁上的缺陷的新方法,这对于永久磁铁电机性能至关重要. 轻量级并行注意剩余网络 (LPAR-Net) 的准确率达到93.63%,超过了工业缺陷识别的现有模型.

关键词:
注意力机制注意力机制深度学习是一种深度学习.机器视觉 机器视觉 机器视觉金属表面缺陷 金属表面缺陷

更多相关视频

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
04:23

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

1.8K
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

2.7K

相关实验视频

Last Updated: Jun 12, 2025

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

485
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
04:23

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

1.8K
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

2.7K

科学领域:

  • 材料科学 材料科学 材料科学
  • 计算机视觉 计算机视觉
  • 工业自动化 工业自动化

背景情况:

  • 永磁电机在工业生产中至关重要,磁质量直接影响其性能.
  • 由于图像特征不清晰,检测磁上的小型和反射表面缺陷具有挑战性.

研究的目的:

  • 开发一种有效的检测磁表面缺陷的方法,解决不清晰的特征和小缺陷大小的挑战.
  • 为了提高工业环境中磁缺陷的识别精度.

主要方法:

  • 使用线性变异来增强细节特征的图像处理.
  • 开发注意力并行剩余卷积块 (APR) 和轻量级并行注意力剩余网络 (LPAR-Net),结合MobileNetV2的倒置瓶结构.
  • 在APR区块中集成了改进的频道注意模块 (CBAM),用于增强特征提取,包括7x7卷积,用于更广泛的空间特征捕获.

主要成果:

  • 拟议的LPAR-Net在磁缺陷数据集上实现了93.63%的准确性.
  • 与主流图像分类模型 (如DenseNet,MobileNet和ConvNext) 相比,LPAR-Net表现优越.
  • 在钢带表面缺陷数据集上的验证证实了该方法强大的识别能力.

结论:

  • LPAR-Net 方法显著提高了磁缺陷检测的准确性和能力.
  • 这种方法为涉及永磁电机的工业制造过程中的质量控制提供了有前途的解决方案.
  • 该研究为表面缺陷分析提供了一个新的数据集和一个经过验证的深度学习模型.