Real Time Face Mask Detection by using CNN
Amruta Chachere, Snehlata Dongre
Abstract
The ongoing Covid-19 (coronavirus) outbreak is worsening the worldwide health crisis, impacting our daily lives. Wearing a face mask is the most basic form of prevention against coronavirus and also it is considered as one of the most significant survival suggestions. Nowadays, the correctness of wearing face masks is manually monitored, and it is impossible to alert people in overcrowded areas or public locations. For this reason, machine learning frameworks such as openCV, keras, scikit-learn, and tensorflow are employed. The proposed approach intends to develop a new way to automatically detect the correctness of face mask in human face. If no facemask is detected, the proposed model will inform or alert the concerned person. To detect face, an openCV with haar-cascade classifier is employed. The Convolutional Neural Network (CNN) model is also used to detect or train the proposed dataset, which includes the images of different persons with or without face mask. This technique leverages an accuracy of about 99.1%.