TY - GEN
T1 - Eyes only
T2 - navigating hypertext with gaze
AU - Penkar, Abdul Moiz
AU - Lutteroth, Christof
AU - Weber, Gerald
PY - 2013
Y1 - 2013
N2 - Eye gaze tracking is an obvious candidate for a future input device, perhaps even for everyday computing. The hard problems with gaze-controlled interfaces are inaccuracy and inadvertent clicking. We attempt to mitigate these problems in the context of a gaze-controlled web browser. Four click alternatives (Dwell, Single Confirm, Multiple Confirm and Radial Confirm) were implemented along with a fifth mouse-controlled version for comparison. Two alternatives make use of additional buttons that confirm a selection made previously by dwell, hence improving accuracy. Our results indicate that the Multiple Confirm alternative performed best among the gaze-based alternatives; it makes use of multiple confirmation buttons when letting the user choose between different options. When compared to the mouse, the clicking times were worse but the accuracy was indistinguishable. User feedback also indicates that, although mouse was considered best, Multiple Confirm was not perceived as slow and generated excitement. This indicates that the Multiple Confirm click alternative has potential as an interaction method for gaze interfaces.
AB - Eye gaze tracking is an obvious candidate for a future input device, perhaps even for everyday computing. The hard problems with gaze-controlled interfaces are inaccuracy and inadvertent clicking. We attempt to mitigate these problems in the context of a gaze-controlled web browser. Four click alternatives (Dwell, Single Confirm, Multiple Confirm and Radial Confirm) were implemented along with a fifth mouse-controlled version for comparison. Two alternatives make use of additional buttons that confirm a selection made previously by dwell, hence improving accuracy. Our results indicate that the Multiple Confirm alternative performed best among the gaze-based alternatives; it makes use of multiple confirmation buttons when letting the user choose between different options. When compared to the mouse, the clicking times were worse but the accuracy was indistinguishable. User feedback also indicates that, although mouse was considered best, Multiple Confirm was not perceived as slow and generated excitement. This indicates that the Multiple Confirm click alternative has potential as an interaction method for gaze interfaces.
UR - http://dx.doi.org/10.1007/978-3-642-40480-1_10
U2 - 10.1007/978-3-642-40480-1_10
DO - 10.1007/978-3-642-40480-1_10
M3 - Chapter in a published conference proceeding
SN - 9783642404795
T3 - Lecture Notes in Computer Science
SP - 153
EP - 169
BT - Human Computer Interaction - INTERACT 2013
A2 - Kotze, P.
A2 - Marsden, G.
A2 - Lindgaard, G.
A2 - Wesson, J.
A2 - Winckler, M.
PB - Springer
CY - London, U. K.
ER -