Robust hand tracking in realtime using a single head-mounted RGB camera
In this paper novel 2D-hand tracking algorithms used in a system for hand gesture interaction are presented. New types of head-mounted Augmented-Reality devices offer the possibility to visualize digital content in the user's field of view. To interact with these head-mounted devices hand gestures are an intuitive modality. Generally, the recognition of hand gestures consists of two main steps: The first one is hand tracking and the second step gesture recognition. This paper concentrates on the first step: Hand tracking. Due to the wearing comfort of the glasses-like systems these only use a single camera to capture the field of view of the user. Therefore new algorithms for hand tracking without depth data are presented and compared to state-of-the-art algorithms by utilizing a thorough evaluation methodology for comparing trajectories.