UTAR Institutional Repository

Real-time face mask detection using YOLOv4

Tey, Chen Hup (2022) Real-time face mask detection using YOLOv4. Final Year Project, UTAR.

[img]
Preview
PDF
Download (8Mb) | Preview

    Abstract

    Coronavirus disease (COVID-19) has been widespread in the world. Starting from December 2019, COVID-19 is a major public health and economic problem because the virus has impacted detrimentally the life quality of people. Therefore, the government has applied a new restriction to force people to wear a face mask to control the spread of coronavirus. However, it is very difficult to manually check if everyone in the public areas is wearing a mask correctly or not. Furthermore, normally the detection results will not be saved or recorded if manually monitoring the people whether they wear masks properly or not. Even though there are many existing face mask detection applications in the market, but most of them do not connect to the database to store the detection results. However, it is important to use those data to analyze the face mask detection result so that important information such as the face mask compliance rate for a specific date can be known. Therefore, this project, which is an automated and powerful real-time face mask detection application that used the YOLOv4 model, has been developed. In this application, the system will apply the multithreading approach to perform real-time detection on two cameras simultaneously. This can save time and human effort in monitoring whether people are wearing masks correctly, as it is inefficient to check whether people wear masks correctly or not manually using pure human effort, especially in public areas. This system will also alert the person by displaying an alarm sound when the person is not wearing a face mask, or the face mask is wearing in an incorrect manner. At the same time, the system will also automatically record the detection results as a video for the user to check or watch the detection results later. Besides that, this application also allows the user to record the detection results manually by taking a screenshot and saving the screenshot into the MySQL database and local drive. Moreover, this application is also connected to the MySQL database to store the detection results. This application will apply an object tracking algorithm when performing face mask detection in real-time to prevent the detections on the same object from being stored into the database repeatedly in every single frame. Furthermore, this application also allows the user to create a face mask detection dashboard for a specific date by using the data stored in the MySQL database for analysis purposes. There is also a function that allows the user to export the detection result from the MySQL database to a CSV file.

    Item Type: Final Year Project / Dissertation / Thesis (Final Year Project)
    Subjects: Q Science > Q Science (General)
    T Technology > T Technology (General)
    Divisions: Faculty of Information and Communication Technology > Bachelor of Computer Science (Honours)
    Depositing User: ML Main Library
    Date Deposited: 13 Oct 2022 15:24
    Last Modified: 13 Oct 2022 15:24
    URI: http://eprints.utar.edu.my/id/eprint/4674

    Actions (login required)

    View Item