Abstract In virtual and augmented reality, gaze-based methods have been explored for decades as effective user interfaces for hands-free interaction. Though several well-known gaze-based methods exist for simple interactions such as selection, no solutions exist for 3D manipulation tasks requiring a higher degree of freedom (DoF). In this paper, we introduce OrthoGaze, a novel user interface that allows users to intuitively manipulate the three-dimensional position of a virtual object using only their eye or head gaze. Our approach makes use of three selectable, orthogonal planes, where each plane not only helps guide the user’s gaze in an arbitrary virtual space, but also allows for 2-DoF manipulations of object position. To evaluate our method, we conducted two user studies involving aiming and docking tasks in virtual reality to evaluate the fundamental characteristics of sustained gaze aiming and to determine which type of gaze-based control performs best when combined with OrthoGaze. Results showed that eye gaze was more accurate than head gaze for sustained aiming. Additionally, eye and head gaze-based control for 3D manipulations achieved 78% and 96% performance, respectively, in comparison with a hand-held controller. Subjective results also suggest that gaze-based manipulation can comprehensively cause more fatigue than controller-based. From the experimental results, we expect OrthoGaze to become an effective method for pure hands-free object manipulation in head-mounted displays.

中文翻译：

---

OrthoGaze：使用正交平面的基于注视的三维对象操作

摘要 在虚拟现实和增强现实中，基于注视的方法作为用于免提交互的有效用户界面已经探索了几十年。尽管存在几种众所周知的基于凝视的方法用于简单的交互，例如选择，但不存在需要更高自由度 (DoF) 的 3D 操作任务的解决方案。在本文中，我们介绍了 OrthoGaze，这是一种新颖的用户界面，允许用户仅使用眼睛或头部注视就可以直观地操纵虚拟对象的 3D 位置。我们的方法使用三个可选的正交平面，其中每个平面不仅有助于引导用户在任意虚拟空间中的凝视，而且还允许对对象位置进行 2-DoF 操作。为了评估我们的方法，我们进行了两项用户研究，涉及虚拟现实中的瞄准和对接任务，以评估持续注视瞄准的基本特征，并确定哪种基于注视的控制与 OrthoGaze 结合使用时效果最佳。结果表明，在持续瞄准方面，眼睛注视比头部注视更准确。此外，与手持控制器相比，基于眼睛和头部凝视的 3D 操作控制分别实现了 78% 和 96% 的性能。主观结果还表明，基于注视的操作比基于控制器的操作更容易引起疲劳。从实验结果来看，我们预计 OrthoGaze 将成为头戴式显示器中纯免提对象操作的有效方法。