| Abstrak/Abstract |
According to the World Health Organization, coronavirus has spread throughout the world and has become a pandemic. One method of spreading this virus is by damaging the droplet that comes out of the mouth/nose of an infected person when breathing, talking, or coughing. So one of the health protocols that must be adhered to is wearing a mask in public places. Therefore, to create a safe and uncontrolled environment, in this study, we created a computer vision-based detection system which implemented into the single board computer raspberry pi 4. A monitoring system in the form of a web server will be implemented. When a violation occurs, the system will capture faces not wearing masks and sound an alarm. In this system, we combine Multi-task Cascaded Convolutional. Neural Network (MTCNN) as face detection and proposed a Convolutional Neural Network (CNN) model for the classification stage. The proposed system can help suppress the spread of the coronavirus. For the overall performance of a proposed system, we calculate Average Precision (AP) and Mean Average Precision (MAP). It achieves 83,33% MAP on daytime testing and 73,5% MAP on nighttime testing. The performance is better on daytime testing because there is less light and more noise at nighttime, so detection becomes more difficult.
|
