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

Corrigendum to "HSF1, in association with MORC2, downregulates ArgBP2 via the PRC2 family in gastric cancer cells" [Biochim. Biophys. Acta. Mol. Basis. Dis. 2018 Apr;1864(4 Pt A):1104-1114/PMID: 29339121].

Biochimica et biophysica acta. Molecular basis of disease·2024
Same author

Marine sulfated glycans inhibit the interaction of heparin with S-protein of SARS-CoV-2 Omicron XBB variant.

Glycoconjugate journal·2024
Same author

Formation of brominated and nitrated byproducts during unactivated peroxymonosulfate oxidation of phenol.

Journal of hazardous materials·2024
Same author

MAMLCDA: A Meta-Learning Model for Predicting circRNA-Disease Association Based on MAML Combined With CNN.

IEEE journal of biomedical and health informatics·2024
Same author

Editorial: Heparan sulfate-binding proteins in health and disease.

Frontiers in molecular biosciences·2024
Same author

The heat shock factor 20-HSF4-cellulose synthase A2 module regulates heat stress tolerance in maize.

The Plant cell·2024

相关实验视频

Updated: May 24, 2025

Estimation of Contact Regions Between Hands and Objects During Human Multi-Digit Grasping
09:41

Estimation of Contact Regions Between Hands and Objects During Human Multi-Digit Grasping

Published on: April 21, 2023

1.5K

VMarker-Pro:从虚拟标记器进行概率 3D 人类网格估计

Xiaoxuan Ma, Jiajun Su, Yuan Xu

    IEEE transactions on pattern analysis and machine intelligence
    |March 3, 2025
    PubMed
    概括

    本研究介绍了单眼3D人网估计的虚拟标记器,克服了深度模糊性和形状损失. VMarker和VMarker-Pro从图像中提供准确而强大的人形重建,即使有遮蔽.

    科学领域:

    • 计算机视觉 计算机视觉
    • 3D人体姿势和形状估计 3D人体姿势和形状估计
    • 机器学习 机器学习

    背景情况:

    • 单眼3D人网估计受到深度模糊性和映射复杂性的挑战.
    • 使用3D姿势的现有方法会失去重要的身体形状信息,而移动捕捉系统对现实世界的图像是不切实际的.
    • 需要强大而准确的方法来从单个图像中进行无标记的3D人形重建.

    研究的目的:

    • 引入创新的中间表示,使用虚拟标记来改进3D人网估计.
    • 开发VMarker,用于检测虚拟标记,并从野生图像中重建完整的人类网格.
    • 用VMarker-Pro,一个使用扩散模型的概率框架来提高封闭场景中的稳定性.

    主要方法:

    • 从大规模的运动捕获数据中学习虚拟标记器,以模仿物理标记器.
    • 开发了VMarker,可以从图像中检测虚拟标记,并插入以获得3D人体网格.
    • 提出了VMarker-Pro,这是一个概率框架,具有扩散模型,通过生成多个合理的网格来处理封闭.

    主要成果:

    • 拟议的VMarker和VMarker-Pro方法超过了对三个基准数据集的现有方法.
    • 在SURREAL数据集中观察到显著的改进,该数据集以各种身体形状而闻名.

    更多相关视频

    Author Spotlight: Enhancing Remote Rehabilitation with Virtual Reality and Electromyography
    04:06

    Author Spotlight: Enhancing Remote Rehabilitation with Virtual Reality and Electromyography

    Published on: January 12, 2024

    539
    Measurement of Dynamic Scapular Kinematics Using an Acromion Marker Cluster to Minimize Skin Movement Artifact
    10:07

    Measurement of Dynamic Scapular Kinematics Using an Acromion Marker Cluster to Minimize Skin Movement Artifact

    Published on: February 10, 2015

    19.1K

    相关实验视频

    Last Updated: May 24, 2025

    Estimation of Contact Regions Between Hands and Objects During Human Multi-Digit Grasping
    09:41

    Estimation of Contact Regions Between Hands and Objects During Human Multi-Digit Grasping

    Published on: April 21, 2023

    1.5K
    Author Spotlight: Enhancing Remote Rehabilitation with Virtual Reality and Electromyography
    04:06

    Author Spotlight: Enhancing Remote Rehabilitation with Virtual Reality and Electromyography

    Published on: January 12, 2024

    539
    Measurement of Dynamic Scapular Kinematics Using an Acromion Marker Cluster to Minimize Skin Movement Artifact
    10:07

    Measurement of Dynamic Scapular Kinematics Using an Acromion Marker Cluster to Minimize Skin Movement Artifact

    Published on: February 10, 2015

    19.1K
  • 通过准确地建模数据分布,VMarker-Pro在封闭的场景中表现出卓越的性能.

    • 结论:

    • 虚拟标记器为单眼3D人网估计提供了一个有希望的中间表示.
    • VMarker和VMarker-Pro提供了精确,坚固和形状意识的3D人体从图像上的重建.
    • 概率框架有效地解决了阻塞挑战,推进了无标记3D人形估计的最先进技术.