An Improved Procedure for Face Mask Detection using Convolution Neural Network

The present COVID-19 pandemic scenario where entire world is facing a lot of unforeseen medical challenges, can be somehow controlled by following some medical guidelines and procedures when monitored strictly. Wearing a specifically recommended face mask is one of the easiest ways to prevent it. However, due to negligence by the common people for avoiding face masks, need a meticulous system to find such people and appropriate action may be taken against them. Medical guidelines state that primarily COVID-19 and its variants outspread typically through nose discharge or droplets of saliva by the infected person's coughs or sneezes. Droplets of saliva all around us could infect us and others. So, People need to wear masks as it's one of the methods which can effectively depress the growth of its spreading. Identifying masks over the faces could be done using many methodologies in Computer Science. The primary concern is to work upon identifying masks on people's faces and increasing their recognition accuracy with simple methodology and network. Keep the fact in mind, this paper presents a methodology for the identification of masks on people's faces using Convolution Neural Networks (CNN), for this, a training dataset has been used to generate more augmented images. Afterward, it has been pre-processed by applying CNN. The method has been evaluated using a CNN-based algorithm on the test dataset. The simulation result shows satisfactory performance and accuracy through different curves.