Identifying patterns for human activities of daily living in smart homes

Abstract

The identification of human activities using image and movement sensors belongs to a new service provided by just a few telecommunication companies, in a “fifth-play” package of services. The paper describes some experiments done by the authors when two persons were present simultaneously in the area covered by a homemade ubiquitous system, for example, a smart home. The bi-dimensional images captured by the image sensor in the kitchen area were filtered to ensure privacy protection. They were further used to build and recognize human activity patterns. The obtained results create the prerequisites to develop new e-health and social services for elder and impaired persons.