Awesome Robotics Manipulation · full_paper

3PoinTr: 3D Point Tracks for Robot Manipulation Pretraining from Casual Videos

作者：Adam Hung · 单位：Carnegie Mellon University · 会议/期刊：arXiv · 日期：2026-03-09 · 来源：Low-Level Learning-Based Action Modelling / Learning Strategy / Vision-Grounded Goal Extraction / Visual Trace

三维表征视频规划基础模型感知机器人学习操作
新标签打开 AlphaXiv新标签打开 Paper新标签打开 Code
3PoinTr: 3D Point Tracks for Robot Manipulation Pretraining from Casual Videos figure
AlphaXiv 中文论文页面（可滚动查看）