Using gaze for 3-D direct manipulation interface

Use of 3-D human motion sensing without physical restrictions is the most promising approach to realize seamless coupling between virtual environments and the real world. Motion capturing without any specific markers by computer vision techniques is the most appropriate for such purposes. As the first step, we have developed an avatar motion control by user body postures, and we have applied it to 3D object manipulation in virtual environments[1]. However, the biggest problem here is that the intention of the user can not be fully recognized only by human body postures, and that, sometimes, unintentional human motions cause unintended object manipulations and incorrect selection of the target object. In this paper, we introduce the user’s gaze to control 3D Direct Manipulation User Interface. Using the gaze it becomes possible control the user interface more accurately and efficiently