DAAM: Drop Activity Alert Manager For Elderly

Falls are a leading and serious problem in elderly individuals due to their aging, cognitive, medications, balance problems, vision and other environmental hazards. These falls can cause broken bones, head injuries and even death if left untreated. A fall can result in decreased quality of life and functional independence. Immediate assistance and medical help should be provided to avoid any complications. The system proposed is a fall detection model based on real-time surveillance. The system applies human pose estimation, object detection algorithm and fall detection conditions on the surveillance video. OpenPose is used for human key point detection combined with YOLO for object detection and various fall detection methods. If a fall occurs, an alert notification is sent to their emergency contacts.

I. Introduction

Globally, there were 962 million elderly persons aged 60 and over in 2017. By the year 2050, they are expected to a number nearly 2.1 billion[1]. Falls are the leading cause of unintentional injuries. The World Health Organization (WHO) states that around 30 - 50% of people living in nursing home and institution experience fall and around 40% are recurrent falls. Aging and weakness increase the chances of falls. Individuals of the age 65 years and over experience hospitalization for fall-related injuries which consist of brain injuries, hip fractures and upper limb injuries. Post-fall syndrome also leads to depression, confusion, immobilization which limits the daily activity of the elderly. The environment also plays an important factor in fall-related injuries which include the placement of rugs, narrow stairs, poor lighting, slippery or uneven floor[2].