Comparison Between the Yolov4 and Yolov5 Models in Detecting Faces while Wearing a Mask

Abstract

Object detection based on deep learning has shown good results ever since the Coronavirus or Covid-19 started sweeping the entire world affecting and killing many people. One of the easiest and simplest ways to protect oneself from this virus is to wear a mask. In order to detect whether a person is wearing a mask or not, we propose here two models for detecting face masks. Facial recognition has been difficult, but with the development of deep learning, it has tremendous ability to detect objects, especially in public places. Therefore, it has become necessary for accurate diagnosis to protect people from Covid-19. In order to discover whether a person is wearing a mask or not, we propose a model to detect face masks whether the mask is worn or not, so it was proposed in this research to use two deep learning algorithms, which are YOLOv5 and YOLOv4, which are among the YOLO models that are characterized by accuracy and speed. Compare learning algorithms Deep and finding the difference between them in performance and accuracy, and there is a CNN algorithm that is also important in discovering things and achieved satisfactory results, but we will use the YOLO model. YOLO is an advanced algorithm with fast, real-time detection. As most of the results were reviewed with previous variations of YOLO and CNN, it is worth noting that the YOLO model is the best model in face detection. Face detection is of great importance in various fields, especially in public places, and requires security accuracy in detection. It is known that investigators' statements about images of masks are not very easy. Therefore, training and evaluation on the dataset available on Google Colab for YOLOv5 and YOLOv4 algorithm are conducted in this paper.