Awesome Robotics Manipulation · full_paper

EgoMI: Learning Active Vision and Whole-Body Manipulation from Egocentric Human Demonstrations

作者：Justin Yu, Yide Shentu, Di Wu, Pieter Abbeel, Ken Goldberg, Philipp Wu · 单位：UC Berkeley, 2xdof.ai · 会议/期刊：arXiv · 日期：2025-10-31 · 来源：Bottlenecks / Generalization / SE(3)/SIM(3)-Equivariance/View Changes

移动操作视频规划模仿学习数据采集泛化操作
新标签打开 AlphaXiv新标签打开 Paper新标签打开 Project
EgoMI: Learning Active Vision and Whole-Body Manipulation from Egocentric Human Demonstrations figure
AlphaXiv 中文论文页面（可滚动查看）