We consider the problem of visual imitation learning without human supervision (e.g. kinesthetic teaching or teleoperation), nor access to an interactive reinforcement learning (RL) training environment. We present a geometric perspective to derive solutions to this problem. Specifically, we propose VGS-IL (Visual Geometric Skill Imitation Learning), an end-to-end geometry-parameterized task concept inference method, to infer globally consistent geometric feature association rules from human demonstration video frames. We show that, instead of learning actions from image pixels, learning a geometry-parameterized task concept provides an explainable and invariant representation across demonstrator to imitator under various environmental settings. Moreover, such a task concept representation provides a direct link with geometric vision based controllers (e.g. visual servoing), allowing for efficient mapping of high-level task concepts to low-level robot actions.
@article{arxiv.2003.02768,
title = {A Geometric Perspective on Visual Imitation Learning},
author = {Jun Jin and Laura Petrich and Masood Dehghan and Martin Jagersand},
journal= {arXiv preprint arXiv:2003.02768},
year = {2020}
}