DINOBot: Robot Manipulation via Retrieval and Alignment with Vision Foundation Models 

Authors: Norman Di Palo and Edward Johns
Institution: The Robot Learning Lab at Imperial College London
Published at: ICRA 2024
Paper: arxiv.org/pdf/2402.13181.pdf
Webpage: www.robot-learning.uk/dinobot
Abstract: We propose DINOBot, a novel imitation learning framework for robot manipulation, which leverages the image-level and pixel-level capabilities of features extracted from Vision Transformers trained with DINO. When interacting with a novel object, DINOBot first uses these features to retrieve the most visually similar object experienced during human demonstrations, and then uses this object to align its end-effector with the novel object to enable effective interaction. Through a series of real-world experiments on everyday tasks, we show that exploiting both the image-level and pixel-level properties of vision foundation models enables unprecedented learning efficiency and generalisation.

Наука