A Comparative study for natural reach-and-grasp with/without finger tracking

Abstract

Grasp interactions often involve both hand tracking and finger tracking to drive the virtual hand deformation and evaluate grasping conditions. With the more involvement of psychology into HCI technology, we are seeing more algorithms employing psychological finding. However, the performances and user experiences of these algorithms remain to be further explored. In this paper, a comparative study has been performed under the same grasping conditions between our formerly proposed method for reach-and-grasp tasks which needs only tracking the hand’s 6-dof motions (Method A) and a typical forward-kinematics enabled virtual grasping method which needs both 6-dof hand tracking and a dataglove for finger tracking (Method B). Virtual spheres centered at the origin with different diameters (i.e., 6 cm, 8 cm and 10 cm) were used as the grasping targets. A panel of 12 participants were divided into two groups and employed in the comparative study on task completion time, accuracy and 3 subjective criteria. It is shown from the experimental results that Method A is better than Method B as far as the above 3 aspects were concerned for simple shapes such as spheres. A demo application was developed using both Method A and Method B, and users’ preference evaluation was performed.