Real-time 2D Human Skeleton Monitoring System

Abstract

Human exposure forecast has also made significant progress in recent years. However, the existing data sets are limited in the context of general exposure estimation challenges. Although there seems to be no significant progress in terms of diversity and difficulty in human body models, the COCO Human Pose Model is sufficient with a variety of data. This comprehensive data set was collected using a built - in taxonomy of more than 200K human activities. The collected images captured human activities from a wider perspective than other data sets including various recreation and occupational activities. The aim of this study is to determine the human body joint and joint localization using the COCO Model as 2D on image. © 2019 IEEE.