Face Mask and Social Distancing Detection for Surveillance Systems

There is a huge panic among the people in recent times due to the spread of communicable diseases. People are in close vicinity to one another when in closed spaces like shops, restaurants, classrooms, etc. There is also a cause for worry in workplaces regarding the safety of the workplace. This paper discusses about two models which can be used to detect the distance between people to ensure social distancing and to detect if people are wearing a mask which can be implemented to follow safety measures. To implement these models deep learning techniques are used. For the social distancing model object detection is done to detect humans and this is done through the YOLOv3. For the mask detection model, the MobileNetV2 is the algorithm which is used for classification. This is used to detect if the people are wearing a mask. These two models can be used for the purpose of prevention against widely spreading diseases. For example, if the people of an organization have to request their customers to stay 6 feet apart or wear a mask in cases where the customers are not following the standard safety protocols, the people of the organization should go directly up to them and request for it. This increases the contact between people and at the same time increases the risk factor for the people working in that organization. When these models are implemented, it reduces unnecessary human contact while also ensuring to alert the customers if they break these protocols.