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

相关实验视频

Updated: Jun 15, 2025

3D Planning and Printing of Patient Specific Implants for Reconstruction of Bony Defects
08:15

3D Planning and Printing of Patient Specific Implants for Reconstruction of Bony Defects

Published on: August 4, 2020

6.4K

玩3D人类恢复的游戏

Zhongang Cai, Mingyuan Zhang, Jiawei Ren

    IEEE transactions on pattern analysis and machine intelligence
    |August 27, 2024
    PubMed
    概括

    使用像GTA-V这样的视频游戏生成大规模的3D人类恢复数据集被证明是有效的. 这些带有自动注释的合成数据显著提高了姿势和形状估计的性能,补充了现实世界的数据限制.

    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

    XOV-Action: Towards Generalizable Open-Vocabulary Action Recognition.

    IEEE transactions on pattern analysis and machine intelligence·2026
    Same author

    Higher order Magnus expansion for driven two-level quantum dynamics.

    The Journal of chemical physics·2026
    Same author

    CuDi: Curve Distillation for Efficient and Controllable Exposure Adjustment.

    IEEE transactions on pattern analysis and machine intelligence·2026
    Same author

    The correlation between physical exercise and cognitive function in older adults with cognitive impairment mediated by inhibitory control and possible cognitive processing mechanisms.

    Frontiers in public health·2026
    Same author

    Line-scan phase measuring deflectometry.

    Optics letters·2026
    Same author

    The effects of exercise on generalized anxiety symptoms in adults: a systematic review and meta-analysis.

    BMC sports science, medicine & rehabilitation·2026

    科学领域:

    • 计算机视觉 计算机视觉
    • 机器学习 机器学习
    • 人类姿势和形状估计

    背景情况:

    • 3D人体恢复 (姿势和形状估计) 已经取得了显著的进展.
    • 现有的数据集在规模和多样性上是有限的,因为动作捕捉成本很高.

    研究的目的:

    • 介绍GTA-Human,这是使用GTA-V游戏引擎生成的大规模3D人类数据集.
    • 调查使用游戏数据用于3D人类恢复的有效性和洞察力.

    主要方法:

    • 使用GTA-V生成大规模的人类序列,并自动注释3D基本真相,使用GTA-V.
    • 在GTA-Human数据集上训练有素的基线和高级模型.
    • 对数据规模,域差距和模型灵敏度进行了系统研究.

    主要成果:

    • 一个简单的基线训练在GTA-人类超越了复杂的方法.
    • 合成数据有效地补充了现实世界的室内数据,解决了领域的差距.
    • 数据集规模和数据密度显著影响模型性能.
    • 在GTA-Human中丰富的SMPL参数标签提供了强大的监督.
    • 福利扩展到更大的模型,如CNN和变压器.

    结论:

    • 游戏生成的数据是大规模3D人类恢复的可行和有效资源.
    • 了解域间隙和数据混合策略对于利用合成数据至关重要.
    • GTA-Human数据集为推进3D人类恢复研究提供了一个可扩展的解决方案.

    更多相关视频

    Author Spotlight: Advancing Human Cardiac Anatomy Through Multi-Scale Analysis of Hearts
    04:22

    Author Spotlight: Advancing Human Cardiac Anatomy Through Multi-Scale Analysis of Hearts

    Published on: June 28, 2024

    410
    3D Printing Model of a Patient's Specific Lumbar Vertebra
    07:30

    3D Printing Model of a Patient's Specific Lumbar Vertebra

    Published on: April 14, 2023

    1.8K

    相关实验视频

    Last Updated: Jun 15, 2025

    3D Planning and Printing of Patient Specific Implants for Reconstruction of Bony Defects
    08:15

    3D Planning and Printing of Patient Specific Implants for Reconstruction of Bony Defects

    Published on: August 4, 2020

    6.4K
    Author Spotlight: Advancing Human Cardiac Anatomy Through Multi-Scale Analysis of Hearts
    04:22

    Author Spotlight: Advancing Human Cardiac Anatomy Through Multi-Scale Analysis of Hearts

    Published on: June 28, 2024

    410
    3D Printing Model of a Patient's Specific Lumbar Vertebra
    07:30

    3D Printing Model of a Patient's Specific Lumbar Vertebra

    Published on: April 14, 2023

    1.8K