CNN-BASED COVID-19 FACIAL MASK DETECTION

Abstract

The COVID-19 pandemic has quickly impacted our daily lives, interfering with international trade and travel. It has become commonplace to wear protective face masks. Many public service providers may soon need their clients to correctly wear masks in order to receive their services. As a result, detecting face masks has become an essential responsibility to support worldwide society. Using certain fundamental machine learning tools, such as TensorFlow, Keras, OpenCV, and Scikit-Learn, this study offers a streamlined method for achieving this goal. The suggested technique accurately recognizes the face in the picture before determining whether or not it is wearing a mask. It can also identify a face and a mask in motion as a surveillance task performance. On two distinct datasets, the approach achieves accuracy levels of up to 95.77% and 94.58%, respectively. Using the Sequential Convolutional Neural Network model, we investigate optimal parameter values to accurately detect the existence of masks without generating over-fitting.