View Independent Human Identification by Gait Analysis using Skeletal Data and Dynamic Time Warping

Microsoft Kinect is a motion sensing device that allows us to conveniently capture human body motions and extract skeletal data. For biometric from human gait, Kinect lets us extract more kind and more complicated gait features which help improve identification accuracy. This paper proposes biometric from human gait technique from view independent gait analysis based on the Kinect’s motion tracking ability in 3D coordinate. Dynamic Time Warping is applied to identify users from their walking styles. The selected features include static features e.g. proportion of body parts’ length , height as well as kinematic features e.g. head tilt angles, body tilt angles.