The Application of the Edge-cloud Computing System Based on Reinforcement Learning in Large-scale Mask Recognition

As the epidemic of the COVID-19 continues, many venues need to be installed with recognition equipment at the entrance to recognize whether people wear masks. However, it is not economical to install multiple high-performance devices with a great recognition model only for mask recognition, and not fast to use the combination of camera shooting devices and cloud computing. Therefore, this paper put forward an edge-cloud computing system, whose purpose is to use the poor computing resources of the edge server to share the computing pressure of the cloud server. In other words, the recognition tasks are assigned not only to the cloud device but also to the edge devices with poor computing power and a simple recognition model. The scheduling optimization problem is solved by reinforcement learning method. In the case of insufficient and imperfect data sets, this system still shows high recognition accuracy and fast response speed.