Architectural framework for implementing visual surveillance as a service

The visual surveillance systems are becoming heavier day by day in terms of number of cameras to manage. It poses a challenge to manage such systems provided that most of these systems are still managed manually. To save cost and to increase reliability in response critical applications, the shift of visual surveillance is towards cloud-based solutions. This paper proposes a framework for implementation of a cloud-based visual surveillance system for human activity recognition. System works in both ways; either video captured by CCTV camera can be live-streamed to cloud in real time or stored videos can be sent to cloud for video content analysis (VCA). The paper advocates some of the improvements in practices and laws pertaining to cloud in order to make its applications widespread.