Crowd Detection that Potentially Violate Covid-19 Health Protocol Using Convolutional NeuralNetwork (CNN)

Naf'an, Emil and Islami, Fajrul and Gushelmi, Gushelmi (2021) Crowd Detection that Potentially Violate Covid-19 Health Protocol Using Convolutional NeuralNetwork (CNN). international Conference on Computer Science and Engineering. (Submitted)

[img] Text
Prosiding+1+Crowd+Detection.pdf

Download (1MB)
Official URL/ URL Asal/ URL DOI: DOI: 10.1109/IC2SE52832.2021.9791865

Abstract

Abstract— This study aims to propose a system that can detect crowds that have the potential to violate the Covid-19 health protocol. The criterion used is the distance between each person in the crowd. In this case, the permissible distance between each person is 1 meter. The camera is used to detect the distance. The input from the camera is connected to the laptop. The images obtained are processed using Deep Learning. In this case, the Convolutional Neural Network (CNN) is used. If the distance between each person in the image is less than 1 meter, the color of that person's bounding box will be red, otherwise, it will be green. In this case, testing was carried out on 3 conditions with the number of people varying from 2 to 20 people. The first condition is the condition that the distance between each person is more than 1 meter. The second condition is the condition that the distance between each person is less than 1 meter and some is more than 1 meter. The third condition is the condition that the distance between each person is less than 1 meter. The proposed system achieves average specificity, sensitivity, and accuracies of 91.03, 91.7, and 91.83, respectively. Keywords— Crowd Detection, Covid-19 Health Protocol, Camera, Convolutional Neural Network (CNN)

Item Type: Article
Subjects: 0 Research > Ilmu Komputer > Jaringan Komputer
0 Research > Ilmu Komputer
Divisions/ Fakultas/ Prodi: Fakultas Ilmu Komputer
Depositing User: Ryan Ariadi A.Md
Date Deposited: 05 Feb 2024 08:54
Last Modified: 05 Feb 2024 08:54
URI: http://repository.upiyptk.ac.id/id/eprint/9674

Actions (login required)

View Item View Item