Face Detection based approach to combat with COVID-19

Abstract

The global problem at the moment is COVID-19 caused by corona-virus which led to the worldwide lockdown. However there are still people who are not taking proper precautions and maintaining social-distancing while going out for emergency situations. To ensure this the government is trying to monitor the actions of each and every citizen. It is not realistically possible to monitor the actions of each and every citizen. This problem can be possibly solved by the use of machine learning. In this paper we have presented face detection based technique to combat with COVID-19. At first this technique captures the image and detects the face using viola zones method. Then the detected face images are classified to detect the mask. It actually detects whether every person have used mask or not. If there present anyone without mask then the system starts alarm to inform it to the responsible authority and to alert the public. Here Principal Component Analysis (PCA) has been used to extract the feature from the detected face images and for the detection of mask K-Nearest Neighbor (KNN) classifier has been used.

Mask Detection Using IoT - A Comparative Study of Various Learning Models

“…Several detection systems [5][6][7]10,11,14,16,17] have been proposed and evaluated using some purposely-built datasets (Table 1) that contain some pictures of people with facemask, without facemask or with facemask put incorrectly. The pictures of masked people either (i) correspond to real-world person that are pictured (real-world dataset) or (ii) result from the addition of a mask picture on an existing facial image (simulated dataset).…”

“…The training relies on the RMFRD and Celeb Faces Attributes 2 dataset. In [5], the face mask detection system captures image, extracts features from image based on Principal Component Analysis (PCA), detects the human face using viola zones method and further uses the K-Nearest Neighbor (KNN) classifier.…”