Visual fall detection system in home environments

Abstract

One important aim of intelligent environments is to allow elderly people to stay as long as possible in their home environment. Fall detection is an essential part of intelligent environments since falls are one of the most distinctive factors that threaten the independence of elderly people. The fall detection system proposed in this paper is based on cameras that are equipped with wide angle lenses which can acquire a complete view of the living space. Our approach uses a novel background subtraction technique to detect the human body in the camera image. The novel technique is especially designed for home environments and can adapt even to strong changes in the scene (illumination changes, open/closed doors) very fast. The orientation of the detected human body is then analyzed to detect falls. First results of our feasibility study show that the visual fall detection system is usable in detecting falls in home like environments of adults and elderly people.