Real-Time Fall Detection based on Shape and Motion Features

Teerasak Kroputaponchai and Nikom Suvonvorn, 2010, Real-Time Fall Detection based on Shape and Motion Features, In Proceedings of Fifth International Conference on Knowledge, Information and Creativity Support Systems (KICSS2010), November 25-27, 2010, Chiang Mai, Thailand, p. 287-262.

In this paper, we propose a real-time system for fall detection in the elderly people based on image analysis techniques, which support the automatic monitoring system of the activities of daily living. Our approach is based on human characteristic analysis using shape and motion. The method is divided into 3 steps: motion detection and tracking based on background subtraction with shadow removal; human features extraction, and fall decision. Experimental results show that our system can detect quite accurately the falls.

Related posts: