Video monitoring-based fall detectors
From Capsil Wiki
Introduction
There has been significant interest in the use of video cameras in the home to detect falls in older people. From a technical perspective video camera promise an effective non contact method of falls detection. However there is strong resistance from older people against the deployment of cameras in homes due to privacy concerns. Methods of hiding identities have been developed such as face or whole body bluring. To date no success methods to circumvent the privacy issues have been developed.
Research Projects
Video monitoring systems use camera’s which attempt to detect a fall event based on image-processing algorithms that are designed to identify unusual inactivity, which is more likely to follow an event of fall. The fall detectors under this category are passive in the sense that they generally do not require the user to wear any device. This form of approach remains an active area of research within the academic community focusing design variations in image-processing algorithms and monitoring/transmission systems [1].
The UbiSense project [2], at the Imperial College London, is focused on developing an unobtrusive health-monitoring system for the elderly by using embedded smart vision techniques to detect changes in posture, gait, and activities. In addition the UbiSense system attempts to capture signs of deterioration of the patients by analyzing subtle changes in posture and gait. Privacy issues are address in UbiSense by immediately filtered the images at the device level into blobs, which encapsulate only the shape outline and motion vectors of the subject. Visual images are not stored or transmitted therefore it is not possible to reconstruct the abstracted image back into the original image.
The University of Dundee, has investigated the use of ceiling mounted camera’s which has the advantage of avoiding the problems associated with furniture occlusion [3]. Patterns of inactivity are used to make inferences about health and also to help detect falls. The University of Liverpool have reported the Smart Inactivity Monitor Using Array-Based Detectors (SIMBAD) fall detector based on a low-cost array of infrared detectors to capture low-level image of the resident and then analyzes the subject’s motion to detect a fall event [4]. The IRISYS sensors are designed to provide non-intrusive monitoring due to the low resolution of the images. Falls are detected using a neural model based on the velocity and acceleration of the tracked object. The findings indicated good specificity in terms of low false-alarm rates; however the model could only detect 30% of the emulated falls.
Despite the ability of video-based fall-monitoring systems to automatically detect falls with no user intervention, the fear of intrusion of privacy are extremely prominent in this technology approach. Although a variety of solutions have been developed to ensure privacy people in homes still experience the feeling of “being-watched” thus making the technology unacceptable in many cases.
References
- P. Rajendran, A. Corcoran, B. Kinosian and M. Alwan, Falls, Fall Prevention, and Fall Detection Technologies in Eldercare Technology for Clinical Practitioners, Humana Press, pp 187-202, 2008.
- http://www.doc.ic.ac.uk/vip/ubisense/home/home.html
- C. H. Nait and S. J. McKenna, Activity summarisation and fall detection in a supportive home environment, International Conference on Pattern Recognition (ICPR), August 23-26, Cambridge, UK, 2004.
- A. Sixsmith and N. Johnson, A Smart Sensor to Detect the Falls of the Elderly, Pervasive Computing, Vol. 3, No. 2, April-June, 2004, pp 42-47.
