Gestural Recognition by a FourKinect Module in a CAVE "Le SAS"

Among the various types of interactions in a virtual reality (VR) room, physical gestures such as moving one’s hands are one of the ways to communicate with the virtual environment. It is possible to capture those gestures with sensors, in our case the Microsoft Kinect. In the context of a large CAVE, not one but four Kinects are used to track the user and capture its gestures. However, the question of the detection of specific gestures by this system comes into play as the CAVE requires recognizing the user’s gestures in order to interpret them into an input. In this paper we focus on the gestural recognition ability of our CAVE applied to three specific gestures: Raising the right hand, raising the left hand and short hopping. Moreover, we provide experimental results showing that a Four-Kinect system provides more effective gesture recognition than a single sensor.