Al Ayyubi, Muhammad Shalahuddin (2021) Sistem Absensi New Normal Menggunakan Convolutional Neural Network Untuk Deteksi Masker. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
10311710000005-Undergraduate_Thesis.pdf - Accepted Version Download (2MB) | Preview |
Abstract
Pada masa pandemi virus Covid-19 absensi manual, analog, dan sidik jari sudah tidak lagi relevan. Hal ini dikarenakan sistem absensi tersebut merupakan fasilitas publik yang sering disentuh oleh banyak orang yang dapat menjadi sarana penyebaran virus Covid-19. Salah satu cara penyebaran virus Covid-19 yang dimaksud adalah bersentuhan
secara tidak langsung dengan permukaan yang telah terkontaminasi. Dari permasalahan tersebut maka penelitian ini bertujuan membangun sebuah sistem absensi yang mengimplementasikan protokol kesehatan dengan fitur seperti registrasi mandiri, deteksi suhu tubuh,
pengenalan wajah dan deteksi masker yang menggunakan Convolutional Neural Network. Pada pengujian sistem absensi new normal didapatkan bahwasannya deteksi masker menggunakan Convolutional Neural Network arsitektur MobileNetV2 memiliki nilai rata-rata tertinggi f1- score dan k-fold cross validation sebesar 99,75% pada dataset training. Pada pengujian dataset percobaan arsitektur MobileNetV2 juga mendapat nilai tertinggi dari arsitektur lainnya dengan f1-score sebesar 93,6% dan rata-rata FPS sebesar 10,17. Dengan keseluruhan sistem, proses absensi memakan waktu rata rata 25,11 detik.
====================================================================================================
During the Covid-19 virus pandemic, manual, analogue and
fingerprint attendance are no longer relevant. This is because the attendance system is a public facility that is often touched by many people which can be a means of spreading the Covid-19 virus. One way to spread the Covid 19 virus in question is indirect contact with contaminated surfaces. From these problems, this study aims to build an attendance system that implements health protocols with features such as selfregistration, body temperature detection, face recognition and mask detection using the Convolutional Neural Network. In testing the new normal attendance system, it was found that mask detection using the Convolutional Neural Network architecture MobileNetV2 has the highest average f1-score and k-fold cross
validation of 99.75% on the training dataset. In testing the experimental dataset, the MobileNetV2 architecture also received the highest score from other architectures with an f1-score of 93.6% and an average FPS of 10.17. With the whole system, the attendance process takes an average of 25.11 seconds
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Uncontrolled Keywords: | Absensi, Covid-19, Deteksi masker, Convolution Neural Network(CNN) |
| Subjects: | Q Science > QA Mathematics > QA76.774.A53 Android Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) R Medicine > R Medicine (General) > R858 Deep Learning R Medicine > RA Public aspects of medicine > RA644.C67 COVID-19 (Disease) |
| Divisions: | Faculty of Vocational > 36304-Automation Electronic Engineering |
| Depositing User: | Muhammad Shalahuddin Al Ayyubi |
| Date Deposited: | 26 Aug 2021 06:36 |
| Last Modified: | 08 Nov 2024 02:03 |
| URI: | http://repository.its.ac.id/id/eprint/90276 |
Actions (login required)
 |
View Item |