Behavior Analysis for Aging-in-Place using Similarity Heatmaps
| Autor: | Mohamed Eldib, Hamid Aghajan, Francis Deboeverie, Nyan Bo Bo, Wilfried Philips, Xingzhe Xie |
|---|---|
| Rok vydání: | 2014 |
| Předmět: |
Service (business)
education.field_of_study Technology and Engineering Aging in place Pixel Computer science business.industry Aging-in-Place Population Economic shortage Machine learning computer.software_genre Sitting Preference Low-Resolution Cameras Similarity (network science) Behaviour Analysis Artificial intelligence education business computer Simulation |
| Zdroj: | ICDSC 8th ACM/IEEE international conference on Distributed Smart Cameras, Proceedings |
| DOI: | 10.1145/2659021.2659038 |
| Popis: | The demand for healthcare services for an increasing population of older adults is faced with the shortage of skilled caregivers and a constant increase in healthcare costs. In addition, the strong preference of the elderly to live independently has been driving much research on "ambient-assisted living" (AAL) systems to support aging-in-place. In this paper, we propose to employ a low-resolution image sensor network for behavior analysis of a home occupant. A network of 10 low-resolution cameras (30x30 pixels) is installed in a service flat of an elderly, based on which the user's mobility tracks are extracted using a maximum likelihood tracker. We propose a novel measure to find similar patterns of behavior between each pair of days from the user's detected positions, based on heatmaps and Earth mover's distance (EMD). Then, we use an exemplar-based approach to identify sleeping, eating, and sitting activities, and walking patterns of the elderly user for two weeks of real-life recordings. The proposed system achieves an overall accuracy of about 94%. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|