Considerations for Targets in 3D Pointing Experiments

We identify various tradeoffs in the design of 3D pointing experiments based on Fitts’ law and the ISO9241-9 methodology. The advantages and disadvantages of several approaches for such experiments are analyzed and compared against each other. Results of an experiment that investigates various visual aids are presented as evidence. We conclude with recommendations for 3D pointing experiments and avenues of future work. Author Keywords Fitts’ law; 3D pointing; ISO 9241-9 ACM Classification Keywords H.5.2. User Interfaces: Evaluation/methodology. INTRODUCTION Pointing at three-dimensional objects to select them is a fundamental task in 3D user interfaces and is analogous to 2D pointing in graphical user interfaces. However, 3D selection is complicated by a number of issues not found in 2D systems. First, 3D graphics systems use perspective; much like reality, far objects appear smaller, which may influence pointing task difficulty. Second, 3D systems often use stereo display to enhance depth perception. Third, there is no universally accepted 3D pointing technique or device. Moreover, there are many different selection methods for 3D targets. Some of the most popular ones require that the users hand or finger intersects the target in space, “laser pointer” techniques, where the user’s hand/finger or a device shoots a virtual ray into the scene and the system then selects the first object along that ray, or touch-through, where the user touches the 2D projection of the 3D object on a touch screen. PREVIOUS WORK Point selection (or pointing) at targets in 2D has been thoroughly investigated using Fitts’ law [16] and the ISO 9241-9 standard [20]. The ISO standard contributes the measure of throughput, which permits comparisons across user strategies by normalizing experimental error rates via a post-experimental accuracy adjustment. This is a primary advantage of the metric – it helps ensure consistency between studies and hence comparability. In contrast, 3D point selection (sometimes simply “selection”) is comparably less well-studied. Numerous 3D pointing studies have been conducted [4, 12, 22, 28, 36], yet 3D pointing is still not as well understood as its 2D equivalent [32]. Most direct 3D selection techniques fall roughly into two broad paradigms: ray-based techniques (including occlusion) and virtual hand techniques [1, 7, 15, 29]. Virtual hands use intersection of the hand/cursor with the target and thus require depth precision. Our current work focuses primarily on techniques requiring depth motions, and hence we mostly focus on virtual hand-based techniques here. Note that these not necessarily use the actual hand (or necessitate finger tracking). For example, our recent work used a tracked stylus to approximate the actual hand position. Virtual hand (and ray-based) selection techniques are largely equivalent to pointing tasks, i.e., they specify a unique position (of an object) in the environment. While many 3D pointing studies have been performed in the past, we focus here mostly on those based on ISO 9241-9. A study of 3D pointing in a stereo display system [7] revealed that a 2D cursor based method was fastest, followed by a pen-based method and then a ray-based method. A recent investigation of cursor-based techniques for 3D pointing also found that methods that 2D cursor-based techniques perform best [38]. Such techniques include the mouse as well as ray-casting with a cursor at the position where the ray hits the screen or the scene. One of the noteworthy outcomes of this work is a 2D model that accounts for perspective distortion, which describes cursor-based pointing at 3D targets quite well. Moreover, displaying the cursor only to the dominant eye was found to significantly improve performance in a stereo display system, even for classic ray casting. A recent comparison of touch-based methods on stereo surfaces revealed that 2D touch, i.e. touch-through, works well for objects within approximately 10cm of the surface, but that 3D touch is better for objects that further from the display [11]. Research comparing pointing in the real world vs. virtual reality indicates that VR performance is substantially worse Paste the appropriate copyright/license statement here. ACM now supports three different publication options: • ACM copyright: ACM holds the copyright on the work. This is the historical approach. • License: The author(s) retain copyright, but ACM receives an exclusive publication license. • Open Access: The author(s) wish to pay for the work to be open access. The additional fee must be paid to ACM. This text field is large enough to hold the appropriate release statement assuming it is single-spaced in TimesNewRoman 8 point font. Please do not change or modify the size of this text box. Every submission will be assigned their own unique DOI string to be