Distributed Person of Interest Tracking In Camera Networks
K. A. Shiva Kumar, K.R. Ramakrishnan, G. N. Rathna · 2017
In this paper we address the problem of tracking a specific person among many in non-overlapping camera networks. It is an important problem with many applications in surveillance and security. Given a person to be tracked by initialising the person in one of the cameras, our objective is to track the person throughout the camera network. We propose a distributed solution where the processing is done in cameras itself and information is shared with neighbours to cooperatively work together. When the person goes out of a field of view (FOV) of a camera, the neighbour cameras go into search mode. We use an on-line fine-tuned re-identification and distributed minimum algorithms to successfully re-identify the person. We performed real-world data experiments showing the performance of the proposed algorithm.