Detection Tracking and Recognition of Human Poses for a Real Time Spatial Game

In this paper, we present an approach to detect, track people, and recognize poses. The detected poses are used for controlling a real time spatial game. In the people detection, tracking and pose recognition system, body parts such as the torso and the hands are segmented from the whole body and tracked over time. The 2D coordinates of these body parts are used as the input of a pose recognition system. By transferring distance and angles between the torso center and the hands into classifier feature space, simple classifiers, such as the nearest mean classifier, are sufficient for recognizing predefined key poses. The output of the classifier, that is the identification of the pose, is used to control color and actions of the virtual actor. The position of the virtual actor is steered by the detected position of the user in the image.