Automated Video Activity Analysis for Eldercare

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): Our objective is to explore new information technologies to assist the independent living of elderly people and enhance their quality of life at home, while utilizing the time and attention of caregivers and eldercare specialists in the highest efficiency. Our multidisciplinary research team will design a privacy-protecting wireless video sensor network to collect continuous video data about the resident's activities at home. We will develop a statistical activity modeling framework to analyze the video data for automated functional assessment and abnormal activity detection. Based on the results from the automated activity analysis, we will develop an video summarization scheme to transform the voluminous video data into a compact and meaningful activity record for fast and efficient functional assessment and problem diagnosis. The abnormal activity detection will help reducing the risks of independent living by detecting falls and other unusual activity patterns that indicate changing conditions or health crisis. The continuous video monitoring and automated activity analysis and summarization will significantly improve the eldercare efficiency. The proposed technologies will be tested and evaluated at TigerPlace, a unique independent-living, apartment-style eldercare facility designed by the Sinclair School of Nursing at University of Missouri-Columbia (MU). Access to this cutting edge facility affiliated with MU provides an exceptional opportunity to accomplish our research aims. The potential risks of the elderly people living alone and the dreadful consequences of late response to abnormal situations call for the proposed technology. The remarkable growth of the elderly population and the shortage of eldercare professionals underscore the significance of the proposed research.