FallStop — A deep learning approach to real-time fall detection and monitoring of vital parameters

Falls can cause bone breakage, immobility and can have a negative impact on the victim's life. The elderly are prone to falls as their eyesight, hearing and reflexes tend to weaken with age. Falls among this population is a growing problem that can be prevented. Many times these falls result in death, in order to avoid this, immediate medical aid needs to be provided to the fall victim. This paper proposes a novel fall detection algorithm and a system for tracking daily movement as well as the vitals of an individual. In addition to this an alert message is sent to the victims emergency contact if a fall or any abnormality in the vitals of the individual is detected. This is performed using You Only Look Once (YOLO) object detection algorithm, Openpose, machine vision and internet of things(IOT).