PHUMA: Physically-Grounded Humanoid Locomotion Dataset

ICLR 2026 Conference Submission10482 Authors

Keywords: humanoid, retarget, dataset

Abstract:

Motion imitation is a promising approach for humanoid locomotion, enabling agents to acquire humanlike behaviors. Existing methods typically rely on high-quality motion capture datasets such as AMASS, but these are scarce and expensive, limiting scalability and diversity. Recent studies attempt to scale data collection by converting large-scale internet videos, exemplified by Humanoid-X. However, they often introduce physical artifacts such as floating, penetration, and foot skating, which hinder stable imitation.
In response, we introduce \textbf{PHUMA}, a \textbf{P}hysically-grounded \textbf{HUMA}noid locomotion dataset that leverages human video at scale, while addressing physical artifacts through careful data curation and physics-constrained retargeting.
PHUMA enforces joint limits, ensures ground contact, and eliminates foot skating, producing motions that are both large-scale and physically reliable.
We evaluated PHUMA in two sets of conditions: (i) imitation of unseen motion from self-recorded test videos and (ii) path following with pelvis-only guidance. In both cases, PHUMA-trained policies outperform Humanoid-X and AMASS, achieving significant gains in imitating diverse motions. PHUMA will be publicly released to support future research.

Primary Area: datasets and benchmarks

Submission Number: 10482