Integration Of Open CV LBF Model To Detect Masks In Health Protocol Surveillance Systems

Abstract

The Corona Viruses Diseases pandemic that was rife in early 2020 and hit many countries caused discipline to be applied to health protocols. The prevention of physical contact between humans gave rise to new traditions in aspects of human life. Almost all public facilities in Indonesia require visitors to wear masks as a means of preventing exposure to viruses in the air. However, this advice is often ignored by some people. In addition to endangering many people, this condition also makes public facility managers need extra resources in the form of time, energy and costs to ensure this health protocol is implemented. The existence of these problems triggers the emergence of innovations to present a system that provides assurance and convenience in ensuring compliance with health protocols for the use of masks through creative and effective methods. This method is done by utilizing CCTV cameras or webcams at the entrance equipped with an Artificial Intelligence program designed to be able to detect the use of masks on visitors to public facilities, and without the need for other sensors. The detection system is built on the concept of facial biometrics and utilizes the OpenCV LBF model to detector landmarks on a person's face. Based on tests conducted through several scenario, it can be said that the open CV LBF model successfully identified the use of masks within 35 seconds, increasing the reading distance to 2 meters making the process longer. In addition, in indoor lighting conditions, the system experienced 1 detection error with a process time of 18 seconds, while for well-light outdoor conditions the system managed to detect all objects within 10 seconds.