Open AccessDeteksi Masker Pencegahan Covid19 Menggunakan Convolutional Neural Network Berbasis Android
Author(s) -
Nyoman Purnama,
Putu Kusuma Negara
Publication year - 2021
Publication title -
jurnal resti (rekayasa sistem dan teknologi informasi)
Language(s) - English
Resource type - Journals
ISSN - 2580-0760
DOI - 10.29207/resti.v5i3.3103
Subject(s) - convolutional neural network , computer science , android (operating system) , artificial intelligence , artificial neural network , covid-19 , deep learning , machine learning , computer security , infectious disease (medical specialty) , disease , medicine , pathology , operating system
Masks are an important part of preventing Covid19 disease.The World Health Organization (WHO) have also recommended the community use masks when doing activities in public areas. There are many types of masks that are used to cover the nose and mouth. In general, there are about 3 types of masks that are commonly used by the public today, namely medical masks, N95 and cloth masks. This study aims to detect the type of mask used by the community. So that it can make easier for the government to apply discipline in COVID-19 health protocol. The detection method used in this study is a convolutional neural network (CNN). The first step is acquisition of knowledge, which first collects the types of masks on the market, followed by the representation of that knowledge before being modeled into a mathematical calculation formula, which will then be processed using the Convolutional Neural Network method. The system will be carried out by analyzing the recall value, its precision and accuracy.Testing process is carried out on an Android-based device and the mobilenetV2 framework. In this study, the accuracy value is 90% using ADAM Optimization and 80 % using Gradient descent optimization.