Abstract
This article presents feature extraction with deep learning for in-hand manipulation. It is important that robot hands can manipulate different sized and shaped objects. In order to generate versatile manipulation with such objects, we used deep learning to extract critical information from object manipulating motions.
