A Fall Risk Prediction System Based on 3D Space Human Skeleton Torso Images

This paper proposes a fall risk prediction system, which adopts two cameras and an AI server to obtain the images of the 3D human skeleton torso activities. We use the Hidden Markov model (HMM) for the movement changes of the joint points of the human skeleton torso in 3D space to evaluate the probability of falling of subjects. The experimental results show that the proposed system can efficiently predict the risk of falling for consumers with unbalanced gait.