iCub World: Friendly Robots Help Building Good Vision Data-Sets
This provides a new dataset for object recognition in robotics, but it is incremental as it focuses on a specific lab setting.
The paper introduces the iCub World dataset, an object recognition dataset acquired using human-robot interaction with the iCub robot, enabling rapid data collection and annotation with ground truth. It demonstrates that this dataset poses challenges to current recognition systems, though it is constrained to a robotics lab environment.
In this paper we present and start analyzing the iCub World data-set, an object recognition data-set, we acquired using a Human-Robot Interaction (HRI) scheme and the iCub humanoid robot platform. Our set up allows for rapid acquisition and annotation of data with corresponding ground truth. While more constrained in its scopes -- the iCub world is essentially a robotics research lab -- we demonstrate how the proposed data-set poses challenges to current recognition systems. The iCubWorld data-set is publicly available. The data-set can be downloaded from: http://www.iit.it/en/projects/data-sets.html.