Face Mask detection using Convolutional Neural Network

Abstract

The ever expanding research with advancement in the field of computer vision provided an innovative solution to face mask detection. An outbreak of infectious disease coronavirus causes severe acute respiratory syndrome. The pandemic diseases at initial stages included the symptoms of cough, fever, dizziness, shortness of breath and fatigue. Although being highly contagious (spread or transmission) this disease has a low rate of mortality with around 80% experiencing a mild effect and 15-20% as high/severe effects, there are no vaccines or specific antiviral medicine available yet but few are in initial stages. Therefore, face mask detectors have become a very important problem in image processing and computer vision. Several recent algorithms have been designed using convolutional architectures to make the algorithm as precise as possible. In this approach, convolutional neural network architecture is applied to design a face mask detector that can detect face in the frame and then label it as “with mask” or “without a mask” The experiments were performed and reached a validation precision of 93.55 after model training.