Development of a face detection system with or without a mask using software

Abstract

This article presents a software with artificial vision using the Python language in Visual Studio Code which is capable of detecting if a person is wearing a mask or not, this project was made in order to be able to identify people who are without masks or misplaced either at work or at the entrance of a premises, for example this would be applied in companies that work with toxic waste, gases, dust and any agent that is transmitted by air, thanks to this it will be possible to identify certain people and give them a feedback on the correct use of the mask. For the development of this software, the main modules such as opencv-contrib-python, numpy and mediapipe were used, in addition to having their dependencies which are installed automatically, as well as a camera with which to capture the images; The first step was to search for a data set and resize it, followed by training it using opencv and finally using mediapipe for face detection and implementing the model in programming. Finally, after the tests carried out, a result of 83.13% validation was obtained, however, its percentage would increase or decrease depending on the quality of the camera, since the errors would decrease if you have a camera with a good quality implemented.