Human Activities Detection for Patient Convalescence

Abstract

A simple activity recognition method can allow a solitary human being to monitor all the surroundings with the purpose to guarantee safety and confidentiality while protective maintenance cost and efficiency with the rising level of accuracy. This monitoring system with real-time video surveillance can be deployed for patients and the elderly in a hospital or old age home and airport along with numerous human activities. For speedy analysis of action and accurate result while working with complex human behavior, we decided to use YOLOv4 (You Only Look Once) algorithm which is the latest and the fastest among them all. This technique uses bounding boxes to highlight the action. In this case, we have collected 4,674 number of dissimilar data from the hospital with different condition of ourselves. During this study, we divided the human action into three different patterns such as standing, sitting and walking. This model is able to detect and recognize numerous patients and other various human activities. This research accomplishes an average accuracy of 94.6667% while recognizing images and about 63.00% while recognizing activity from video clips. This study works with YOLOv4 while it performs better than TensorFlow and OpenPose platforms. The article proposed the outcome for patients in early recovery based on human activities investigation and analysis.