Human motion tracking and feature extraction for cognitive rehabilitation in informationally structured space

This paper discusses measurement methods of human motions based on 3D distance image sensor, and human interaction of rehabilitation using robot partners. We focus on rehabilitation programs for Unilateral Spatial Neglect. First, we explain robot partners and sensor networks for rehabilitation support. Next, we apply a method of extracting human motions from 3D distance image by using growing neural gas based on distance criteria. Furthermore, we propose a human motion analysis method based on evolution strategy and neural network. Finally, we discuss the effectiveness of the proposed methods through several experimental results.