HiPHI：面向高精度人体运动与物体交互的大规模基准数据集
HiPHI: A Large-Scale Benchmark for High-Precision Human Motion and Object Interaction - Wiley Science and Engineering Content Hub
HiPHI 是一个 617.5 小时的全身体人体运动数据集，通过光学动捕以亚毫米精度采集，其中包含 245.7 小时带同步物体轨迹与网格的人-物交互数据，并用 FrameNet 组织动作覆盖范围。该白皮书还提出用于衡量运动多样性与交互 grounding 的基准套件，并报告了在该数据集上训练的策略部署到实体 Unitree G1 人形机器人上的结果。
正文
当前显示原文，中文全文译文尚未就绪。
White Paper
Access
First register your details to create a user profile for the hub, then login to access all content within the hub.
Already registered? Click here to log in to the hub.
Login information
Personal profile
Download this complimentary White Paper today!
This White Paper gives robotics researchers and engineers an overview of a new large-scale motion capture dataset built to close the data gap limiting humanoid robot learning. It also shows how policies trained on the dataset transfer to a real humanoid robot.
What you will learn about:
Why humanoid robot learning, a central problem in embodied AI and Physical AI, needs data that internet video and existing motion capture datasets cannot provide; How FrameNet, a linguistic framework for human action, can guide motion capture collection to systematically cover a broad range of whole-body motion; Why synchronized object trajectories and meshes make human-object interaction data useful for teaching robots real-world tasks such as carrying, pushing, and pulling; How reinforcement learning policies trained on this motion capture data improve with scale, and how sim-to-real transfer carries them onto a physical humanoid robot
Click ‘LOOK INSIDE’ to Download Now.
More Information
Humanoid robots must learn to balance, move, and interact with objects across an enormous range of situations, but the data available for this training has clear limits. Internet video shows diverse behavior but cannot capture precise physical states. Laboratory motion capture systems record accurate movement but usually cover only a narrow set of actions. This white paper examines HiPHI, a 617.5-hour whole-body human motion dataset captured with optical motion capture at sub-millimeter accuracy. It includes 245.7 hours of human-object interaction with synchronized object trajectories and meshes, and organizes coverage using FrameNet, a linguistic framework for human action. The paper also introduces a benchmark suite for measuring motion diversity and interaction grounding, and reports results from policies trained on the dataset and deployed on a physical Unitree G1 humanoid robot.
Access
First register your details to create a user profile for the hub, then login to access all content within the hub.
Already registered? Click here to log in to the hub.
Noitom Robotics builds ModalityNet, the human-centric data substrate that makes the physical world learnable for embodied AI.
Partnered with
IEEE Spectrum Magazine, the flagship publication of the IEEE, explores the development, applications and implications of new technologies. It anticipates trends in engineering, science, and technology, and provides a forum for understanding, discussion and leadership in these areas.
1. Same tier 3 topics + same sponsor [1], 2. Same tier 2 topics + same sponsor [1], 3. Any topic + same sponsor [1], 2. Same tags + same sponsor [1], 3. Same sponsor [1],
1. Same tier 3 topics + same partner [3],
Tagged With; benchmark dataset; embodied AI; FrameNet; human motion dataset; human-object interaction dataset; human–object interaction; humanoid robot learning; motion tracking; motion-space coverage; object trajectory data. Supplementary keywords (client-provided): motion capture; optical motion capture; Physical AI; reinforcement learning; robot policy learning; sim-to-real transfer; sub-millimeter tracking; whole-body control; whole-body motion