SMART HUMAN ACTIVITY DETECTION USING YOLOv4

Smart Human Activity Detection using YOLOv4 is a commanding research area in computer vision due to its numerous applications such as person surveillance, social distancing detection, object detection, etc. A simple activity recognition model can allow a single human person to monitor all our surroundings to ensure safety and privacy while preserving maintenance cost and efficiency with a high level of precision. This technique uses bounding boxes to highlight the action. The monitoring system with real-time video surveillance can deploy for patients and the elderly in a hospital or nursing home and various human activities in essential areas such as the airport. For speedy analysis of action and accurate results, while working with complex human behavior, we decided to use the YOLOv4 (You Only Look Once) algorithm, the latest and fastest among them all. This model detected and recognized multiple patients or other regular person activity and multiple human activities tracing support at once. To see human activity, a CNN model has been pre-trained. The accuracy of convolutional neural networks is typically 79-90% when used to detect actions by adding pooling and convolution layers, followed by a few fully connected layers and recognizing patterns in an interval. It is estimated to provide an accuracy of 79-90%. Applications include surveillance and observation with the proposed system. This paper helps detect moderate human activity using YOLOv4 and Darknet techniques.