DirectionQ: Continuous Mid-air Hand Input for Selecting Multiple Targets through Directional Visual Cues

Abstract

We propose DirectionQ, an interaction technique using a continuous mid-air hand input that shows directional visual cues on objects to enable fast and accurate multiple target selection in dense environments. The user can select various targets in a row by moving their hand consecutively in the same direction with the visual cues of targeted objects. DirectionQ involves assigning directional visual cues to each object in the user's viewpoint, detecting the trajectory changes of hand movement, and inferring intended targets with directions of the segmented hand movement. Unlike traditional approaches for multiple target selection, DirectionQ does not need additional steps to disambiguate, providing less time for multiple target selection. While distant, small, or occluded objects often bring accuracy or time-consuming problems, DirectionQ shows similar performance regardless of target conditions. To prove the effectiveness of our approach, we conducted two controlled experiments comparing our system with two existing techniques, RayCasting and EyeGaze. The results show that DirectionQ is faster and more accurate than two techniques with a lower cognitive load on multiple target selection.