4D Visual Pre-training for Robot Learning

Chengkai Hou, Yanjie Ze, Yankai Fu +5

General visual representations learned from web-scale datasets for robotics have achieved great success in recent years, enabling data-efficient robot learning on manipulation task…