Privacy-Preserving Recognition of Activities in Daily Living from Multi-view Silhouettes and RFID-based Training.
Sangho Park, Henry Kautz · 2008
There is an increasing need for the development of sup-portive technology for elderly people living indepen-dently in their own homes, as the percentage of elderly people grows. A crucial issue is resolving conflict-ing goals of providing a technology-assisted safer en-vironment and maintaining the users ’ privacy. We ad-dress the issue of recognizing ordinary household activ-ities of daily living (ADLs) by combining multi-view computer-vision based silhouette mosaic and radio-frequency identification (RFID)-based direct sensors. Multiple sites in our smart home testbed are covered by synchronized cameras with different imaging reso-lutions. Training behavior models without costly man-ual labeling is achieved by using RFID sensing. A hierarchical recognition scheme is proposed for build-ing a dynamic Bayesian network (DBN) that encom-passes various sensing modalities. Advantages of the proposed approach include robust segmentation of ob-jects, view-independent tracking and representation of objects and persons in 3D space, efficient handling of occlusion, and the recognition of human activity with-out exposing the actual appearance of the inhabitants. Experimental evaluation shows that recognition accu-racy using multi-view silhouette mosaic representation is comparable with the baseline recognition accuracy using RFID-based sensors.