• cloud
• fog testbed
• video streaming
• neural networks
• people counting
• web application
• fog computing
• iot

We present an indoor monitoring system prototype capable of monitoring the flow of people in a building and estimating the number of persons inside. The prototype exploits a set of IoT devices that elaborate cameras images in real-time. The prototype also includes a collector unit, that gathers data produced by the devices, and provides an interface to visualize aggregated information and video-feedback for debug purposes. In terms of provided functionalities, the prototype is capable of estimating the number of persons in a building, monitoring the crowding levels at the entrances, and detecting whether persons entering the building (do not) wear facial masks.