Design and evaluation of 3D selection techniques based on progressive refinement

Issues such as hand and tracker jitter negatively affect user performance with 3D selection techniques based on the ray-casting metaphor. This makes it difficult for users to select objects that have a small visible area, since small targets require high levels of precision. We introduce an approach to address this issue that uses progressive refinement of the set of selectable objects to reduce the required precision of the task. We present three exemplar techniques (Sphere-casting refined by QUAD menu (SQUAD), Discrete zoom, and Continuous zoom) and derive a design space for progressive refinement from their characteristics. We explore the tradeoffs between progressive refinement and immediate selection techniques in two studies: first comparing SQUAD to ray-casting; and second comparing the zooming techniques to raycasting. In both studies, an analytical evaluation based on a distal pointing model and an empirical evaluation demonstrate that progressive refinement selection can provide significant benefits compared to immediate techniques. In the first study, SQUAD was much more accurate than ray-casting, and SQUAD was faster than ray-casting with small targets and less cluttered environments. The issue with SQUAD, however, is that it requires all selectable objects to be visually distinct. The zooming techniques address this issue by exploring other areas of the progressive refinement design space. They allow users to use the spatial relationships among objects as criteria ∗Corresponding author. Tel.: +1 540 235 6969; fax: +1 540 231 9218 Email addresses: [email protected] (Felipe Bacim), [email protected] (Regis Kopper), [email protected] (Doug Bowman) Preprint submitted to International Journal of Human-Computer StudiesOctober 27, 2011 for selection and to increase precision without requiring precision in pointing. The results of the second study show that while the zooming techniques were significantly slower than ray-casting, accuracy was much higher. Additionally, depending on the size of the target, users chose not to use zoom and, therefore, performed as fast as with ray-casting.