Color-based object tracking in multi-camera environments
Katja Nummiaro, Esther Koller-Meier, Tomáš Svoboda, Daniel Roth, Luc Van Gool · 2003
Abstract. Smart rooms provide challenging research fields for surveillance, human-computer interfacing, video conferencing, industrial monitoring or service and training applications. This paper presents our efforts toward building such an intelligent environment with a calibrated multi camera setup. For a virtual classroom application, the best view is selected automatically from different cameras that follow the target. Real-time object tracking, which is needed to achieve this goal, is implemented by means of color-based particle filtering. The use of multiple models for a target results in a robust tracking even when the view on the target changes considerably like from the front to the back. Information is shared between the cameras and used for (re)initializing an object with epipolar geometry and the prior knowledge of the target model characteristics. Experiments in our research environment show the possible uses of the proposed system. 1