A Neck-Floor Distance Analysis-Based Fall Detection System Using Deep Camera

Abstract

The research content of this paper is to present a skeleton analysis-based system for senile people, which can detect the fall accident. Depth image is applied to extract the 3D joint coordinate of the senile people. In the proposed algorithm, three points on the ground are selected to get the plane equation of the ground plane. The 3D coordinate of the neck is tracked continually, and the space length from the neck to the ground plane is analyzed. If the joint of neck is close to the floor and this situation lasts for more than one min, fall is detected and alert is sent to the relatives or healthcare centers. In this study, images of living room are taken to built a data set, and the proposed method is compared with shape analysis method and deep learning methods. The proposed method gives a better result.