Abstract
In this paper, we present an interactive augmented reality (AR) game for tracking a remote-controlled car controlled by players. We propose it as a new markerless framework for tracking a colored remote-controlled car by integrating a Bayesian classifier into particle filters. This adds the useful abilities of automatic track initiation and recovery from tracking failures in a cluttered background. A Bayesian classifier is utilized to determine the car‘s color probability before tracking. In addition, by using the online adaptation of color probabilities, this method is able to cope well with luminance changes. We calculate the projection matrix as an online process. The method presented can be used to develop the real-time game of AR to remote-controlled car playing. The application can entertain players interactively by controlling the car to the augmented items. A user study is conducted to evaluate the effectiveness of the aforementioned application.