Pengembangan Sistem Pendeteksi Masker Sesuai Protokol Kesehatan dengan Algoritma Mobilenetv2 dan Raspberry Pi

Abstract

A new type of human coronavirus was discovered in December 2019 in Wuhan, China. In humans, coronaviruses usually cause respiratory tract infections, ranging from the common cold to serious diseases such as Middle East Respiratory (MERS) and Severe Acute Respiratory Syndrome (SARS). This new type of coronavirus was later named Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-COV2) and caused Coronavirus Disease-2019 (COVID-19). COVID-19 can cause mild to severe symptoms. So, wearing a mask and keeping a distance is very important to stop the spread of COVID-19. In previous research, a deep learning model has been developed to identify whether the person is wearing a mask or not. In previous studies, the classification was limited to whether humans wore masks or not. There is no classification as to whether the use of masks is right or wrong and whether the masks worn are masks that are in accordance with the recommendations of the Ministry of Health. So that in this study, the detection system for the use of masks is able to detect the use of masks in accordance with the recommendations of the Indonesian Ministry of Health which refers to the interim WHO Guidelines June 5, 2020, regarding recommendations regarding the use of masks in the context of COVID-19, namely the use of cloth masks, medical masks, and masks. can ensure the cover of the mouth and nose, and adjust to the bridge of the nose. The result is a system with the SSDLite Mobilenet V2 model has the highest FPS compared to a system using a system with SSDMNV2. That is, the maximum FPS obtained is 3.57 FPS and the minimum FPS is 3.45 FPS