Unobtrusive fall detection using 3D images of a gaming console: Concept and first results

Conference: Technik für ein selbstbestimmtes Leben - 5. Deutscher AAL-Kongress
01/24/2011 - 01/25/2012 at Berlin

Proceedings: Technik für ein selbstbestimmtes Leben

Pages: 6Language: englishTyp: PDF

Personal VDE Members are entitled to a 10% discount on this title

Marzahl, Christian; Penndorf, Peter; Staemmler, Martin (Fachhochschule Stralsund, Stralsund, Deutschland)
Bruder, Ilvio (Universität Rostock, Institut für Informatik, Rostock, Deutschland)

Image based fall detection is demanding, costly and is rated obtrusive by those being monitored. The approach presented in this paper uses a cost efficient gaming console for 3D image generation. The image itself covers a range of about up to 30cm above the floor and allows for a nearly invisible positioning e.g. under the bed. Image analysis results in the classification of events like "feet in front of the bed", "fall", "leaving the room" and "activity in the room". For usage in nursing homes and in home environments a system design has been implemented which is compatible with the guidelines of the Continua Health Alliance. Events and alarms are forwarded to the carers using a mobile phone or the available nurse alarm system. In addition, the system supports the nursing home in its obligations for documentation of events and eases answering the requests of insurers and in liability issues. The implementation is compliant to data security requirements by transmitting image data to a central server solely on demand for documentation of an event. The system has successfully been tested in a laboratory environment and is currently validated in three rooms of a nursing home in order to prepare for a large scale trial in multiple nursing homes.