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

Reconstruction of plantar degloving injuries using a prefabricated anterolateral thigh plantar glabrous skin flap.

Journal of orthopaedic surgery (Hong Kong)·2026
Same author

Research on Tin Bath Bottom Bricks for Float Glass Furnaces.

Materials (Basel, Switzerland)·2026
Same author

Study on the Synergistic Enhancement of Mechanical Properties of Magnesia-Chrome Refractory Bricks Through Component Ratio Optimization and Salt Impregnation Process.

Materials (Basel, Switzerland)·2026
Same author

Sucrose Unloading in Grains Modulates the Disparity in Grain Filling of Superior and Inferior Spikelets in Rice.

Physiologia plantarum·2026
Same author

Vaccine Effectiveness and Clinical Characteristics of Breakthrough Varicella During Outbreaks Among Students in Qingdao, China.

The Pediatric infectious disease journal·2026
Same author

Neural network-based practical prescribed time adaptive tracking control for nonlinear networked control systems under deception attacks.

Neural networks : the official journal of the International Neural Network Society·2026

相关实验视频

Updated: May 21, 2025

Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography
11:34

Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography

Published on: May 15, 2017

11.0K

FDTransUnet:一个基于特征差异化的表面缺陷细分模型.

Mingzhu Tang1,2, Wencheng Wang1,2

  • 1College of Mechanical and Control Engineering, Guilin University of Technology, Guilin, China.

PloS one
|March 19, 2025
PubMed
概括

本研究介绍了FDTransUnet,这是工业缺陷细分的新型模型. 它通过使用功能差异化来增强数据和混合U-net变压器架构来提高准确性和稳定性.

科学领域:

  • 计算机视觉 计算机视觉
  • 机器学习 机器学习
  • 材料科学 材料科学 材料科学

背景情况:

  • 工业缺陷细分面临着数据有限,识别精度低,细分精度差的挑战.
  • 现有方法在样本大小不足的情况下扎,导致过度装配和模型性能降低.
  • 准确的缺陷检测对于制造过程中的质量控制至关重要.

研究的目的:

  • 提出FDTransUnet,一种利用特征差异化对的表面缺陷细分模型.
  • 解决数据稀缺问题,提高识别和细分的准确性,增强模型的概括性.
  • 为工业检查场景开发一个强大的解决方案.

主要方法:

  • 采用特征差异化数据增强策略来扩大有限的缺陷样本并减轻过度拟合.
  • 整合变压器架构到U-net中,将全球自我注意力与层次结构相结合,以有效提取信息.
  • 构建了一个复合损失函数来处理前景-后台类不平衡,并提高细分精度.

主要成果:

  • 在表面缺陷数据集上,FDTransUnet实现了94.5%的平均像素精度 (MPA) 和89.7%的子系数.
  • 在钢表面缺陷数据集上的概括实验表明,与主流模型相比,FDTransUnet的表现强.
  • 该模型在不同的工业检查场景中表现出良好的概括性能和稳定性.

更多相关视频

Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns
13:44

Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns

Published on: August 30, 2013

42.6K
Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
05:56

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Published on: April 14, 2023

2.3K

相关实验视频

Last Updated: May 21, 2025

Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography
11:34

Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography

Published on: May 15, 2017

11.0K
Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns
13:44

Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns

Published on: August 30, 2013

42.6K
Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
05:56

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Published on: April 14, 2023

2.3K

结论:

  • FDTransUnet有效地克服了数据限制,并提高了表面缺陷细分的准确性.
  • 混合U-net变压器架构和复合损耗功能有助于提高性能.
  • 拟议的模型显示了对现实世界工业检查应用的巨大潜力.