Visual tracking of objects via rule-based multiple hypothesis tracking

In this paper, one of the most crucial step of a visual surveillance system is presented. To track the multiple objects in the scene, multiple hypothesis tracking is combined with the fuzzy logic. Mixture of Gaussians method has been used to detect the moving objects in the video, which is taken from a static camera. Kalman filter has been utilized to estimate the next state of the objects. After the estimation, current measurements have been compared with the estimated features by utilizing fuzzy rules. The proposed method has been tested for both single and multiple camera configurations.