AI Resonance · 阅读最新日报 · AI 入门推荐

2026-11-18 · America/Los_Angeles · 论文 · #6

TCP4D: A Unified Tool-Center-Point Tracking and Generation Model for Robot Manipulation

Robot manipulation videos provide a rich visual experience, yet recovering robot action and gripper-state annotations from RGB observations still remains challenging. Using the embodiment-agnostic tool-center-point (TCP) pose as the action space, we present TCP4D, a unified model in which a shared 4D scene representation both tracks observed TCP motion and generates future TCP trajectories. Given a monocular video and initial 2D TCP queries, the tracker directly recovers metric TCP poses in SE(3) and gripper states alongside scene geometry and camera parameters. The generator samples local…

TCP4D: A Unified Tool-Center-Point Tracking and Generation Model for Robot Manipulation

热度 47.3 / 100;排名与评分保留该期记录。