An adaptive sensor fusion based objects tracking and human action recognition for interactive virtual environments
Yeong Nam Chae, Young-Ho Kim, Jin Woo Choi, Kyusung Cho, Hyun Suk Yang · 2009
As augmented reality is advanced, more smart and sophisticated virtual environments are required. The purpose of the present work is to enable person in an interactive virtual environment to simultaneously and conveniently interact with virtual agents. In this paper, we propose a realtime system that track people in virtual environment robustly and recognize their action view-invariantly based on adaptive sensor fusion approach between laser scanner and video camera. To track the objects precisely, we match objects of video camera to the datum of laser scanner adaptively. Then we recognize human action view-invariantly based on a motion history image and a moving trajectory of objects. To evaluate the performance of the proposed system, we employed it augmented reality application where users can interact with a virtual pet, named cho-rong-i. The results confirm that reliable tracking is archived and persons' actions can be recognized for interactive virtual environment.