Abstract
Human-friendly robots will work mainly in daily living spaces for humans, such as offices, homes and public spaces. In these environments, there are generally many objects which can cause the visual occlusion. Then it will often occur that the robot cannot see some object by the occlusion while the user can, and vice versa, because they will stand in different positions. In these situations, it is hoped that the robot interacts with the user while recognizing the difference between their fields of view. In this paper, we propose a method for estimation of robot's and user's fields of view from images obtained by a stereo camera. First, in our proposed method, user's head direction is estimated using the particle filter. Second, surrounding environments are recognized and then robot's and user's fields of view are estimated. Experimental results show that our method works effectively.
