A Cognitive IoT Smart Surveillance Framework for Crowd Behavior Analysis


Smart and proactive surveillance using IoT framework has recently attracted a lot of attention due to the impracticality of traditional monitoring to handle large amounts of data from distributed cameras. However, in the case of IoT platforms for video processing, centralized cloud servers may fail due to high latency and high bandwidth usage, leading to the need for distributed processing with fog/edge computing. Moreover, tools and techniques are also necessary to intelligently analyze incoming data with minimum human intervention. This paper proposes a fog/edge computing-based IoT smart surveillance framework that adds additional cognitive knowledge in the form of informative frame selection and attention maps. The experimental analysis was conducted in the context of crowd behavior analysis, and the simulation results proved that the main concerns of latency and bandwidth are addressed efficiently by the proposed framework.



Software And Hardware