Rule-based multiple object tracking

We present in this paper a novel framework for multiple object tracking. The proposed algorithm consists of two main steps. The first one initializes the correspondences between detected objects and tracks by using the spatio-colorimetric model of tracks coupled with the appearance-based descriptors of objects. Two color-based models are used to enhance the robustness of the tracking result. In the case of merging multiple tracks, the second step is introduced to correct the fusion and separate the merged tracks of the compound region. The global system has been tested on different well-known datasets. Experimental results demonstrate the robustness of the proposed system in several difficult scenarios.