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

Microcracking in Concrete01:20

Microcracking in Concrete

87
Microcracking in concrete refers to the tiny cracks that can form within the material even before any external load is applied. These microcracks typically occur at the interface between the coarse aggregate and the hydrated cement paste, often as a result of differential volume changes prompted by variations in stress-strain behavior, as well as thermal and moisture movement. Initially, these microcracks remain stable and do not grow substantially until the concrete is stressed to about 30...
87

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

Effects of red palm oil intervention on serum lipids, hepatic antioxidant capacity and gut microbiota in high-fat diet-fed mice.

Open life sciences·2026
Same author

Portraying Ethical Risks of Medical AI: Mixed Methods Study From Connotation Definition to a Survey on Physicians' Cognition.

Journal of medical Internet research·2026
Same author

Early recurrence of atrial tachyarrhythmia after catheter ablation and its associations with clinical outcomes, atrial fibrillation burden, and blanking period duration: A post hoc analysis of the CABANA trial.

Europace : European pacing, arrhythmias, and cardiac electrophysiology : journal of the working groups on cardiac pacing, arrhythmias, and cardiac cellular electrophysiology of the European Society of Cardiology·2026
Same author

Spatial Epidemiology of the Ischemic Heart Disease-Asthma Comorbidity: A Global Analysis of Burden Patterns, Risk Drivers, and a Composite Risk Index.

Risk management and healthcare policy·2026
Same author

Efficacy and optimal timing of sivelestat in critically ill patients with COVID-19: a multicenter retrospective cohort study.

BMC pulmonary medicine·2026
Same author

Karacoline attenuates sepsis-induced acute lung injury by suppressing apoptosis via PPARγ-associated inhibition of JNK/ERK MAPK signaling.

Respiratory research·2026

相关实验视频

Updated: May 10, 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

430

裂-MsCGA:一个深度学习网络,为路面裂检测提供多层次的注意力.

Guoxi Liu1, Xiaojing Wu1, Fei Dai1

  • 1College of Big Data and Intelligent Engineering, Southwest Forestry University, Kunming 650224, China.

Sensors (Basel, Switzerland)
|April 26, 2025
PubMed
概括

这项研究介绍了Crack-MsCGA,这是一种用于路面裂检测的深度学习网络. 它通过避免低级特征融合和使用多尺度注意力来提高小裂的精度.

关键词:
多个尺度的注意力融合.多尺度裂检测多尺度裂检测路面裂检测 路面裂检测小规模的裂检测小规模的裂检测.

更多相关视频

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

319
Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
06:37

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

2.4K

相关实验视频

Last Updated: May 10, 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

430
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

319
Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
06:37

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

2.4K

科学领域:

  • 土木工程 土木工程是指土木工程.
  • 计算机科学 计算机科学
  • 人工智能的人工智能

背景情况:

  • 路面裂检测对于道路安全和维护至关重要.
  • 目前的卷积神经网络 (CNN) 方法由于低级特征融合干扰而难以处理小裂.
  • 现有的注意力机制增强了全球特征,但无法完全解决局部细节损失.

研究的目的:

  • 开发一个计算效率高的深度学习网络,用于准确的多尺度路面裂检测.
  • 改进小规模裂的检测,这些裂具有微妙的局部结构和不同的全球形态.
  • 通过避免低级特征融合来减少噪音干扰.

主要方法:

  • 提出了一个新的深度学习网络,Crack-MsCGA,利用CNN和多层次的注意力.
  • 引入了多层次关注机制 (MsCGA),专注于地方和全球信息的高层特征.
  • 实现了对短距离依赖的局部窗口注意力和对长距离依赖的级联小组注意力,通过混合局部通道注意力 (MLCA) 融合.

主要成果:

  • 在DH807数据集上,Crack-MsCGA在各种尺度上提高了检测准确度.
  • 在AP@50中显著改善:小规模破裂为11.3%,中型破裂为8.1%,大规模破裂为5.9%.
  • 在路面裂检测方面表现优于现有的五种最先进的方法.

结论:

  • 拟议的Crack-MsCGA网络有效地解决了现有多尺度路面裂检测方法的局限性.
  • 避免低级特征融合,并采用一种新的多尺度注意力机制,可以提高检测准确度,特别是对于小裂.
  • 该方法提供了一个有前途的解决方案,通过精确的裂识别来提高道路安全性和优化维护.