Identifying People Using Temporal and Spatial Changes in Local Movements Measured from Body Sway

We propose a novel method of identifying people using temporal and spatial changes in local movements measured from a video sequence of body sway. Existing methods identify people using a gait feature mainly representing the large swinging of the limbs. The use of the gait feature introduces a problem in that the identification performance decreases when people stop walking. To extract an informative feature from people who have stopped walking, our method measures small swings of the body, which is called body sway. We extract the feature from local movements of body sway by participially dividing the body into regions. Experimental results for a dataset of body sway of 118 participants show that the local movement feature obtained using our method outperforms the gait feature obtained using an existing method.