Both-hand gesture recognition based on KOMSM with volume subspaces for robot teleoperation

Abstract

This paper implements a real-time hand gesture recognition system capable of reliably discriminate a potentially large number of hand gestures using both hands to teleoperate a robot. For that we make use of Kernel Orthogonal Mutual Subspace Method (KOMSM, an extension of MSM) and introduce the definition of Volume Subspace, which is the type of subspace generated by using depth images from sensors like Microsoft Kinect. We propose to take advantage of the properties of Volume Subspace to accurately classify hand gestures for robot teleoperation.