Towards Compressive Camera Networks

The scale and scope of multi-camera networks has advanced significantly; camera networks are now found not only in surveillance and access control applications but also in motion capture systems, light stages for reflectance acquisition, and large scale traffic monitoring. While the specifics of the individual applications differ, certain broader trends transcend these applications; examples of such trends relate to the scalability of the network in terms of the increasing number of cameras and the increasing capabilities of individual cameras. In this article, we study the role of compressive sensing (CS) in multi-camera networks and its central role in enabling scalability of network. Specifically, we focus on the central question of whether recent advances in CS and sparse representations can enable solutions to scalability challenges in large-scale multi-camera networks. Some of this discussion is speculative, focused on the potential opportunities afforded by fundamentally re-architecting camera networks by exploiting CS based approaches to tackle the data deluge challenge inherent in multi-camera networks.