CORE4D: A 4D Human-Object-Human Interaction Dataset for Collaborative Object REarrangement

Yun Liu; Chengwen Zhang; Ruofan Xing; Bingda Tang; Bowen Yang; Li Yi

doi:10.1109/cvpr52734.2025.00172

CVPR2025顶会

CORE4D: A 4D Human-Object-Human Interaction Dataset for Collaborative Object REarrangement

Yun Liu, Chengwen Zhang, Ruofan Xing, Bingda Tang, Bowen Yang, Li Yi

arXiv DOI 出版方

2025年份

20顶会引用

摘要

github.io/ Figure 1. CORE4D is a large-scale diverse human-object-human interaction dataset for collaborative object rearrangement, encompassing real-world and synthetic branches. CORE4D-Real captures 1K human-object-human mesh sequences with allocentric and egocentric visual signals, while CORE4D-Synthetic retargets real-world data onto 3K virtual object shapes formulating 10K motion sequences.