Fog Assisted Visitor Identification Framework with Improved Latency and Network Usage

Visitor identification is one of the vital problems of society. The visitor identification process consists of an intelligent security camera that monitors the entrance of the door of a residence, taking pictures of the human when they enter, automated techniques extract the human face from the image and identify them with the help of a database in real-time. However, there is no such existing research that can deal with visitor identification using fog computing. In this paper, a visitor identification framework is proposed that is entirely deployed in the fog computing environment. The proposed prototype is deployed in the Java-based iFogSim simulator. Each simulation uses a different number of cameras for video stream generations. The human object part is trimmed and selected from the raw data for processing. The trimmed data is transferred to the fog node instead of the cloud for processing. This paper deals with how the generated data pass through each module and takes how much amount of time to process the data in distinct modules. Generally, the fog-based framework reduces latency, network usage, and energy consumption. The research finds an improvement in overall latency and network usage of fog computing environments over the cloud environment.