Person Re-identification in Wide Distributed Camera network

Person re-identification in a wide area requires attention to the topology of cameras and the design an efficient re-identification process, which results in the correct matching in the shortest time and the least data traffic. In this paper, a person re-identification method in a wide area is proposed based on the DNPR method and the cluster tree topology. The matching cameras are divided into clusters and there is a head cluster for each cluster. During a hierarchical process, re-identification operations are carried out in two phases of learning and matching. The simulation results show that the proposed method has a higher efficiency in terms of first-rank recognition rate, matching time and traffic within the network than the DNPR method.