Imagine2Real: 通过视频生成先验实现零样本人形物体交互

Jiahe Chen, ZiRui Wang, Feiyu Jia, Xiao Chen, Xiaojie Niu, Weishuai Zeng, Tianfan Xue, Xiaowei Zhou, Jiangmiao Pang, Jingbo Wang

2026

摘要

本文提出Imagine2Real框架，通过将机器人和物体运动统一为4D点轨迹，并利用关键点跟踪器仅追踪稀疏关键点，避免了重定向过程。该方法利用行为基础模型的潜在空间作为搜索域，结合渐进式训练策略，实现了零样本物理部署。

humanoid-object interactionzero-shotvideo generative priors4D point trajectoriesbehavior foundation model

MANIPULATION

📊 2 引用

Yusen Li, Ziwei Wang, Xiangye Zhu 等 12 位作者

Robotics and Computer-Integrated Manufacturing · 2026

MANIPULATION

📊 1 引用

Shaoqing Qin, Lida Zhu, Yanpeng Hao 等 10 位作者

Robotics and Computer-Integrated Manufacturing · 2026

MANIPULATION

📊 0 引用

Bo Li, Yuanbo Zhao, Huijie Xiao 等 6 位作者

Robotics and Computer-Integrated Manufacturing · 2026

MANIPULATION

📊 0 引用

Hao Li, Yuhui Yu, Rui Fu 等 6 位作者

Robotics and Computer-Integrated Manufacturing · 2026