Real time COVID-19 facemask detection using deep learning
Prasad N Achyutha, Swaroop Hebbale, V Vani
Abstract
The COVID-19 pandemic is producing a global health pandemic. According to the World Health Organization (WHO), the utmost effective protection is to wear a face mask in crowded regions/areas. During this pandemic, it is compulsory for every person to wear a mask and maintain social distancing. In the field of Image Processing, Convolutional Neural Networks (CNNs) have risen to prominence as the most common type of image realization/recognition model. Our project's purpose is to research and assess Machine Learning (ML) technologies for identification and recognition of people wearing face masks in any pre-recorded videos, photos, or in actual-time (real- time) circumstances. Our project aims to create a real-time Graphics User Interface based Automated Facial Recognition as well as Mask Detection System. The algorithms used in the proposed methodology are Principal Component Analysis (PCA) and the HAAR Cascade Algorithm. Finally, the result is indicated by a “GREEN” color rectangle box, which would be drawn around the section of the face, which indicates that the person on the camera is wearing a mask, or a “RED” color rectangle box, which indicates that the person on the camera is not wearing a mask. This model achieves 99% accuracy.