LAUSR

This paper proposes a novel network flow model for multi-target tracking, which uses short and highly reliable detection responses as the basic unit, namely the tracklet, in the model. Our model exploits the local information of the tracklet and deploys the global strategy of data association in tracking. The method is divided into two phases: a local phase and a global phase. In the local phase, our method is used to track targets using the detection results, namely the tracking by detection, where the boosted particle filter is used to generate high-confidence detection responses and they are connected into reliable tracklets. In the global phase, the multi-object tracking is modeled as data association problem, and the problem is represented by the maximum posterior probability. Finally, the model is solved by the minimum cost flow algorithm. When dealing with the target occlusion, this paper designs a two-step optimization algorithm to solve the long-term occlusion that affects tracking. A large number of experimental results show that our method is more effective in multi-object tracking than other state-of-the-art methods.