AR Study: Gaze Fastest, Finger Best for Central Vision Loss

Researchers from the Human-Computer Interaction lab (Chen et al.) conducted a study with 20 people with low vision (PLV) and 18 sighted controls to evaluate three object selection techniques in augmented reality: head pointing, gaze pointing, and finger pointing – all using dwell-based confirmation. The study mimicked two real-world scenarios: sitting and walking (on the go). Key results showed that for PLV, gaze-based selection enabled the fastest initial pointing when seated, and overall selection time was comparable to head-based selection in both scenarios. However, due to reduced gaze stability among PLV participants, head-based selection proved to be the most stable and least mentally demanding technique overall.

Interestingly, participants with central vision loss uniquely preferred finger-based selection, reporting a greater sense of control. This preference was not observed in those with other types of low vision or in sighted controls. The study provides empirical insights into accessible AR interaction techniques for selection-based vision enhancements, suggesting that no single method works for all – instead, adaptive interfaces that offer multiple input modalities could better serve the diverse needs of PLV users in augmented reality environments.