Identify pointing and waving in gesture based Human-Machine-Interaction
The gestures "pointing" and "waving" are investigated with regard to their characteristic properties and their suitability for using in gesture based Human-Machine-Interaction (HMI). The MS Kinect 2® as usable motion capturing device is proofed. A description of different basic types of gestures in human-to-human interaction is given and the requirements for the HMI are discussed. A general phase model in combination with the movements of a gesture execution is explained. For the technical recognition of these movement sequences, parameters are defined which are based on the joint data of the Kinect. Detecting the pointing gesture an angle, its angular velocity and optional a holding time are used. The waving gesture is also detected with the help of an angle and its periodicy. To evaluate these angles and the necessary threshold values, several experiments had been done.
Rechteinhaber: ©2017 - TU Ilmenau
Nutzung und Vervielfältigung:
Alle Rechte vorbehalten