MF131 Development of the system to detect falling behavior
Yoshitaro YAMANAKA, Atsushi ODA, Naoto Tojo, Hajime Terasaki · The Proceedings of the JSME Symposium on Welfare Engineering · 2007
Authors are now considering the system that offers various services by monitoring daily life operations of a client. And as first application, we are developing the system which detects falling down behavior and report it. The system, which including a stereo camera in the room of client, extracts a person area from the image using a background subtraction method. Secondly, the system compares time series data about the person area with time series data about falling down behavior which already studied, in advance by HMM and then, judges the client has fallen down or not. In this paper, we report the result of evaluating the performance of this experimental system that is set up in an actual client's room.