PhD forum: Flexible clustering in smart camera networks

Abstract

Clustering is an important concept for structuring large networks of cameras. In this research we investigate the various trade offs of clustering in networks of smart cameras. Major research questions in this context are (i) how to model clustering, (ii) how to deal with the heterogeneity in large camera networks, and (iii) how to integrate clustering in real-world networks. We have developed a flexible and scalable software suite that supports clustering in camera networks. We present first results in a multi-camera person tracking case study.