A System for Recommending Essential Visual Perception Information in Life Monitoring Using Air Pollution Data

Abstract

Life monitoring systems with air pollution data have been developed for solitary elderly people. However, the existing systems obtain unnecessary information that invades their privacy. In addition, since degrees of air pollution change in distinctive ways due to the effect of home environment and airflow condition, it is difficult for people who have no expert knowledge to identify activities of their daily living (ADL). To solve the problems above, this paper presents a system to select useful information that is required to identify ADL. Experimental results have shown that the proposed system achieves 77.3% identification accuracy. Besides, compared with the case when 10 users were asked to identify ADL without the proposed system, the identification accuracy has increased by 11.3% with the proposed system.