Cutting-Edge Deep Learning Approaches for Human Pose Estimation and Activity Recognition in Smart Cities

Human Pose Estimation and Activity Recognition using deep learning are critical components for rising smart cities. These technologies influence convolutional neural networks (CNNs) and recurrent neural networks (RNNs) to accurately perceive and track human postures and actions from video data. By automating the analysis of human behavior, they enhance urban security, improve traffic management, and enable intelligent healthcare monitoring. Various systems into smart city infrastructure can lead to safer, more responsive, and efficient urban spaces. In this paper, focus on implement an architecture which can automatically identify the human activity that is being performed in the video with a high range of accuracy. We have also used Media Pipe and OpenCV to implement a pose estimation model which can multiply use cases in health and sports.