Single camera based 3D tracking for outdoor fall detection toward smart healthcare
Myeongseob Ko, Suneung Kim, Kyungchai Lee, Mingi Kim, Kwangtaek Kim · 2017
Detection of falling is an important research issue to develop smart healthcare technologies for elderly people as the population is rapidly increasing. To achieve the goal, accurate human tracking is imperative to detect a fall regardless of distance and occlusions. However, existing technologies are focusing on either 2D tracking with a single camera or 3D tracking using a depth camera (distance limitation) or multiple cameras (high complexity). In this paper, we introduce a new approach that combines Extended Kalman Filter (EKF) based 2D image tracking and 3D depth tracking to improve the performance of fall detection with a single camera. The experimental results demonstrate that our approach is a promising way of automatic fall detection that becomes a core technology of a smart healthcare solution for the elderly.