DMCA
Learning Appearance Features to Support Robotic Manipulation (2002)
Cached
Download Links
| Venue: | Proceedings of the Cognitive Vision Workshop, electronically published |
| Citations: | 9 - 0 self |
BibTeX
@INPROCEEDINGS{Piater02learningappearance,
author = {Justus H. Piater and Roderic A. Grupen},
title = {Learning Appearance Features to Support Robotic Manipulation},
booktitle = {Proceedings of the Cognitive Vision Workshop, electronically published},
year = {2002}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
A predominant purpose of vision is to facilitate interaction with the world. Motivated by biological vision, we argue that many common vision-guided activities can be carried out directly on the basis of features of appearance, and do not require elaborate world models. We describe a visual feature learning system that supports a haptically guided, dextrous robotic grasping system. It learns features, combinations of Gaussian-derivative filter responses, that correlate well with successful grasping parameters. Without explicit knowledge of object identities or categories, the system learns to propose object-specific grasp parameters, considerably improving the quality of haptically guided grasps with respect to the "blind" system. The combined system is loosely anthropomorphic in that it is guided by vision for hand pre-shaping, and by haptics during execution of a grasp, without explicit object recognition, scene reconstruction, or path planning.
Keyphrases
support robotic manipulation appearance feature blind system path planning dextrous robotic grasping system elaborate world model many common vision-guided activity explicit object recognition object-specific grasp parameter explicit knowledge predominant purpose biological vision gaussian-derivative filter response object identity visual feature scene reconstruction combined system hand pre-shaping
