John Vladimir G. Son, Jose Rico L. Suan, Christian V. Maderazo
Although it is becoming more and more advocated, the use of face masks by the general population to stop the COVID-19 epidemic is contentious, and its potential is not fully known. Despite the complexity of the issues, there is a widespread belief that there is not enough evidence to warrant the use of face masks, particularly among the general public in public places. The researchers aim to provide a system that detects a person who is not wearing a facemask or improperly using a facemask. The system makes use of deep learning techniques to detect persons who are improperly using a facemask in a live feed. The data that will be used will be from readily available sources on the internet which collectively has over 2,000 images that fall under the 3 categories With Mask; Without Mask; and Masks worn incorrectly. The researchers will use a Convolutional Neural Network to classify images from the gathered data, the classification will then be used for the K-nearest neighbor algorithm to group the images into their corresponding classes. The researchers hope to develop and deploy a system that detects whether a person is wearing their mask properly or not. The researchers want to evaluate how effective the system is when classifying an image of people with and without face masks. The researchers would like the system to alarm other people in the area with a short subtle sound that has classified a person not wearing a mask or wearing a mask incorrectly. © 2023 IEEE.
University of San Carlos, Department of Computer Information Sciences and Mathematics, Cebu City, Philippines
