Automatic Behavioral Pattern Analysis of the Elderly

10 Apr
Friday, 04/10/2009 5:00am to 7:00am
Ph.D. Dissertation Proposal Defense

Adams Williams

Computer Science Building, Room 150

The growing numbers of elderly individuals in need of support to live in the community will severely test the current services infrastructure. Part of the solution is to develop technology to increase the length of time elders can remain at home by monitoring their health and assisting them with essential daily tasks such as preparing food and medicating. A common measure for determining an elder's health is their ability to perform certain essential daily activities, known as activities of daily living (ADLs), as well as the nature of their general patterns of activity over a period of time. I propose to develop a camera-based distributed sensor network that can be placed in an elder's home in order to monitor their ability to perform ADLs, track changes in activity patterns over different time scales, and detect key anomalies such as if the person has fallen down. This poses significant challenges in visual tracking and time series data modeling that must be addressed. The system will also include a graphical user interface that summarizes the data in a manner that is useful to health care professionals.

Advisor: Allen Hanson