Aplikasi Deteksi Pengguna Masker Menggunakan Convolutional Neural Network

Abstract

Covid 19 virus –can spread and infect humans through droplets that come out of the mouth and nose of people who are infected with this virus. One of the efforts to prevent the transmission of this virus is to apply the applicable Health protocols, especially the use of masks. Various studies have proven the effectiveness of the se of masks in preventing respiratory tract infections such as COVID-19 reaching above 90%.  This study focuses on making Web-based applications that can detect human faces in real time when using a mask, not a mask. The method used in this research is Convolutional Neural Network with MobilenetV2 Architecture. The CNN method with the MobilenetV2 architecture was chosen because this method has good results in classifying 2-dimensional image data  and  the resulting  model  training  results  have  a  99.6%  fairly  light computation. The Confusion Matrix has an accuracy rate of 99.1%.