ABSTRACT
At a time when the coronavirus pandemic is at its peak and a third wave could be right around the corner, the importance of wearing a mask in public spaces cannot be emphasized enough. Existing research shows that wearing a mask in public reduces the spread of infection by 70% yet only 44% of the population reportedly wears a mask in public. Hence, we found it imperative that a real-time system should be designed to detect whether a person is wearing a mask or not. This chapter presents a simplified approach to achieve this purpose using some basic deep learning concepts like the creation of neural networks and data flow graphs (TensorFlow), data preprocessing (Keras, NumPy), to plot the training loss and accuracy (Matplotlib), to detect objects from our camera (OpenCV), and max pooling (MobileNet). Our software can be used in public spaces to catch violators of government safety protocols without requiring human intervention and should attain good accuracy on different datasets.
