Attendance monitoring of masked faces using ResNext-101

Abstract

SARC virus, Coronavirus, Ebola and bird flu have all caused pandemics in the last few decades. Most of these diseases spread through the air when someone coughs, sneezes or even talks. The government makes citizens wear masks. Furthermore, all academic activities are conducted in virtual mode as a result of this predicament, making taking attendance of pupils difficult while they are wearing masks on their faces. To overcome this issue, the proposed work will track a student’s attendance while using ResNext-101 in a virtual classroom setting. ResNext-101, a deep learning technique, is used on masked faces in this work and it is a good model for accurately detecting masked faces. By using the Gaussian data augmentation approach, the outcome reveals a level of accuracy of 51.70 percent with a loss of 1.9452.