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

Identification of chemical constituents of Smilax glabra Roxb. and comparative study on pharmacokinetics of 12 main bioactive components in normal and hyperuricemia rats by UPLC-Q-Exactive Orbitrap-HRMS method.

Journal of pharmaceutical and biomedical analysis·2025
Same author

Divergent Synthesis of Polycyclic N‑Heterocycles via Controllable Cyclization of Divinylallenes Catalyzed by Copper and Gold.

JACS Au·2025
Same author

Impact of exercise interventions on cardiotoxicity in patients with breast cancer undergoing chemotherapy: a study protocol for a quasi-randomised controlled trial.

BMJ open·2025
Same author

Pan‑cancer analysis of the carcinogenic role of WSB2 in human tumors.

Molecular medicine reports·2025
Same author

Tunable Single-Cell Multistate Nanomechanical Phenotyping by Micropipette-Assisted Atomic Force Microscopy to Dissect Cellular Heterogeneity.

ACS applied materials & interfaces·2025
Same author

A single-dose mRNA vaccine encoding the classical swine fever virus E2-ECD induces durable protective immunity in rabbits.

Veterinary research·2025

相关实验视频

Updated: May 6, 2026

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment
08:25

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Published on: May 7, 2019

8.6K

在复杂的交通场景中,数字双胞胎辅助图匹配多任务对象检测方法在复杂的交通场景中.

Mi Li1,2, Chuhui Liu3, Xiaolong Pan4,5

  • 1College of Information Science and Engineering, Jiaxing University, Jiaxing, 314001, China. limi@zjxu.edu.cn.

Scientific reports
|March 29, 2025
PubMed
概括

这项研究引入了一种新的数字双胞胎方法来检测交通对象,生成虚拟数据以克服现实世界的收集挑战. 该方法有效地跨领域转移知识,在各种交通场景中提高检测准确性和稳定性.

关键词:
数字双胞胎 数字双胞胎 数字双胞胎图形匹配的匹配方法多任务对象检测多任务对象检测虚拟数据集是一个虚拟数据集.

更多相关视频

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

1.3K
Utilizing vmTracking to Improve the Accuracy of Multi-Animal Pose Estimation in Rodent Social Behavior Studies
07:34

Utilizing vmTracking to Improve the Accuracy of Multi-Animal Pose Estimation in Rodent Social Behavior Studies

Published on: November 7, 2025

559

相关实验视频

Last Updated: May 6, 2026

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment
08:25

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Published on: May 7, 2019

8.6K
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

1.3K
Utilizing vmTracking to Improve the Accuracy of Multi-Animal Pose Estimation in Rodent Social Behavior Studies
07:34

Utilizing vmTracking to Improve the Accuracy of Multi-Animal Pose Estimation in Rodent Social Behavior Studies

Published on: November 7, 2025

559

科学领域:

  • 计算机视觉 计算机视觉
  • 人工智能的人工智能
  • 运输工程 运输工程

背景情况:

  • 交通数据的收集和注释是耗时和劳动密集的.
  • 当前的深度学习模型在实际交通应用中面临局限性.
  • 虚拟和现实世界数据集之间存在分布差异.

研究的目的:

  • 为交通场景提出使用数字双胞胎的跨域对象检测传输方法.
  • 解决数据收集和模型限制在实际交通分析中的挑战.
  • 提高物体检测在各种交通环境中的稳定性和适应性.

主要方法:

  • 使用模拟平台构建了一个数字双胞胎流量场景,以生成虚拟流量数据集.
  • 开发了一种多任务对象检测算法,利用图形匹配来对准源域和目标域之间的特征分布.
  • 整合了注意力机制,例如细分和多级别的歧视器,用于对抗性培训,以增强代表性学习.

主要成果:

  • 通过全面的比较实验证明了生成的虚拟数据集的实际价值.
  • 验证了拟议的基于图形匹配的传输方法的有效性.
  • 展示了数据集提高任务性能的能力,以及方法的稳定性和适应性.

结论:

  • 数字双胞胎方法为生成大规模的注释交通数据集提供了可行的解决方案.
  • 拟议的跨域传输方法在交通场景中显著提高了对象检测性能.
  • 该方法在各种交通条件下表现出稳健性和适应性,突出其实际适用性.