Object Detection in Video

Abstract

A recent advances in the discovery of face mask have made the object detection a hot topic for research. In the fields of computer vision object detection is the basic but a very tough and difficult task. With its ability to extract the powerful features, essential for the accretion to detect an object. A Deep learning methods has discovered it’s areas of practice in the field over the past few years after that it get the success. We have powerful methods, but the conditions for obtaining a face and decide whether that face has mask or not has their different and complex challenges, it is like real-time detection, changing climate, and complex lighting conditions. In this research project, We have focus Single Object Detection in Video, which aims to confirm whether the person wear the mask or not by using object detection method. In year 2020 a dangerous virus disease named COVID-19 has affected our daily life and negatively affected the communal health, world trade and global economy. As Covid-19 virus spreads through the population, mutations and symptoms will be changed. In march 2021 new delta variant detected. As per the latest news recently new variant named Omicron is detected in Botswana. To contribute to public health, the aims of this paper is to explore a more accurate and real-time approach that can better visualize a non-mask face in public. Therefore, the discovery of a face mask has become an important task of helping the international community. There are extraordinary styles of algorithms to be had, YOLOV4 stands out from all the different gift currently. The custom dataset have been used to understand face masks and have been skilled on those dataset for detection and monitoring. For evaluation of the skilled model, Precision, keep in mind and Accuracy turned into calculated, it really works by using comparing the ground-fact bounding field vs the detected box and, in the long run, returns the accuracy rating. The higher the Accuracy score could be, the better version is within the detection of items.