HUMAN ACTIVITY RECOGNITION AND FALL DETECTION

Abstract

Personalized monitoring and its application is increasing with the advancement of technology. And during pandemics its become very essential to keep an eye on the prone area and one of the area to identify was old age home. Dizziness, unconsciousness, and others are the common problems associated with elderly people due to weakness and this was also the symptoms of covid. So an unusual activity of falling of elderly people was very difficult to identify and also to monitor. The technology was updated till now to identify posture of normal activity such as running, walking, jumping and many but revert to that falling was an area need to explore. During the fall of an elderly person, the injuries are very fatal, and to void this case we proposed a design to identify the fall and try to notify the system about its fall. Although we try to predict the fall so that it becomes easy to monitor and provide medical help as soon as possible. The main theme is to identify the posture activity and once identify we will compare the activity with trained datasets and if it’s normal in vision them no notification occurred and if the percentage of falls was high then we can predict the system as fall video