A distributed camera network architecture supporting video adaptation

Sparked by the desire to secure human presence in high risk areas, this paper proposes a distributed camera network architecture. This architecture is intended as a robust platform for video surveillance systems. In addition, the proposed system has the ability to collect/capture video streams from cameras at a point-of-interest and deliver a consolidated version of the video to a receiver. To achieve this, video sources are able to discover nodes, adapt content, and transmit adapted video streams to a receiver. Our framework takes into account different issues related to the decentralized Peer-to-Peer paradigm such as peers' unreliability, as well as pragmatic aspects like receivers' heterogeneity. Adaptation ensures that video quality is proportional to bandwidth availability as well as the receiving devices' resources. Simulation is used to manifest that the framework is robust, reliable and suitable for multiparticipant real-time collaboration and real-life deployment.