Towards Object Prediction based on Hand Postures for Reach to Grasp Interaction

Daiber Florian, Valkov Dimitar, Steinicke Frank, Hinrichs Klaus, Krüger Antonio

Zusammenfassung

Recently, traditional multi-touch surfaces are extended by stereoscopic displays and 3D tracking technology. While reaching and pointing tasks have a long tradition in human-computer interaction (HCI), the hand pre-shaping which usually accompanies them has rarely been considered. The Reach to Grasp task has been widely investigated by many neuropsychological and robotic research groups over the last few decades. We believe that subtle grasping hand postures in combination with stereoscopic multi-touch displays have the potential to improve multi-touch 3D user interfaces. We present a study that aims to identify if the intended object can be predicted in advance, relying only on detection of the hand posture.