HiPHI: A Large-Scale Benchmark for High-Precision Human Motion and Object-Interaction

2026-08-17

AI Summary

This study addresses the bottleneck in humanoid robot policy learning caused by the mismatch between embodied data precision and coverage. We propose a theory-guided data construction paradigm leveraging the FrameNet linguistic framework to direct optical motion capture. This approach yields a high-fidelity, whole-body motion dataset exceeding 600 hours, featuring sub-millimeter human-robot interaction trajectories and mesh-level object reconstruction, alongside comprehensive evaluation benchmarks. By significantly expanding motion coverage while preserving interaction fidelity, this work effectively reconciles the trade-off between data scale and precision. Consequently, it establishes a scalable, high-quality data foundation essential for robust policy training and generalization in humanoid robotics.

Abstract

Humanoid intelligence requires learning over an extremely diverse space of whole-body motions and physically grounded interactions. However, existing embodied datasets remain fundamentally limited: internet-scale video data lack precise physical states and interaction grounding, while laboratory motion datasets provide high fidelity but only narrow behavioral coverage. This mismatch creates a critical bottleneck for scalable humanoid policy learning. We present HiPHI, a 600+ hour scale high-fidelity whole-body human motion dataset designed to systematically maximize coverage of the human motion and interaction manifold. HiPHI is theoretically guided by FrameNet, a linguistic framework organizing human primitives. Created using an optical motion capture pipeline, HiPHI provides sub-millimeter spatial marker tracking accuracy for full-body human motion and mesh-level object trajectories. We further introduce a benchmark suite evaluating motion-space diversity, interaction grounding, object consistency, and physical AI applications. Our analyses demonstrate that HiPHI significantly expands motion coverage compared to existing motion datasets while maintaining high-fidelity interaction quality, and establishes a scalable data foundation for training, evaluating, and generalizing humanoid policies in real-world embodied tasks, where similar extensions are also applicable to motion prior models in computer graphics.

Problem

• Humanoid Intelligence
• Embodied AI
• Motion Capture Dataset
• Object Interaction
• Policy Learning

Innovation

• High-Fidelity Motion Capture
• FrameNet-Guided Data Curation
• Human-Object Interaction Benchmark
• Whole-Body Motion Manifold
• Sub-Millimeter Tracking Accuracy