People Counting in the Times of Covid-19

ABSTRACT

Estimating the number of people within a public building with multiple entrances is an interesting problem, especially when limitations on building occupancy hold as during the Covid-19 pandemic. In this article, we illustrate the design, prototyping and assessment of an open-source distributed Cloud-IoT service that performs such a task and detects crowd formation via EdgeAI, also accounting for privacy and security concerns. The service is deployed and thoroughly assessed over a low-cost Fog infrastructure, showing an average accuracy of 94%. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.