Modeling abnormal walking of the elderly to predict risk of the falls using Kalman filter and motion estimation approach

It is estimated that in 2050, more than one fifth of the world’s population will be over 65 years old and these people will face serious risks in their future life. In this paper, a new method is proposed which models the motion aggregation pattern by receiving video strings containing the walks of the elderly and tracking their motion to identify the position of the joints involved in movements. Then, drawing on biomechanical laws of motion and joint angle estimates, the skeletal framework is mapped on the image, which is eventually transferred to the 3D space. The data include a set of 322 video strings taken from CASIA and CAVIAR databases and a sample of the falls in the elderly taken from Sabzevar’s Mother Nursing Home in Iran. The results show the effective performance of the algorithm in identifying the risk of falls associated with abnormal walking of the elderly.