XIHAO

Traditionally, in order to perform artificial intelligence functions on edge devices such as IoT sensors and mobile phones, the application or system had to be hosted on a cloud platform. With the exponential increase of edge devices and data generated, capabilities of cloud computing would eventually be limited by the bandwidth and latency of the network. To mitigate the potential risks posed by cloud computing, this report discusses the possibility to move processing of data to where it is being generated by having Artificial Intelligence at the Edge. To evaluate this new computing paradigm, a visitor management system with facial recognition was implemented. The facial recognition model created was able to achieve an accuracy of 94% on static datasets with an inference speed of less than one second per photo while running on a RaspberryPi. The visitor management system was deployed and tested on both edge and cloud architectures. Results shown that performance of the system is..