From 2D to 3D

Mid-air text entry on virtual keyboards suffers from the lack of tactile feedback, which brings challenges to both tap detection and input prediction. In this paper, we explored feasibility single-finger typing QWERTY in mid-air. We first conducted a study examine users' 3D behavior different sizes keyboards. Results showed that participants perceived vertical projection lowest point keyboard during as target location inferring taps based intersection between finger was not applicable. Aiming at challenge, derived novel prediction algorithm took uncertainty into calculation probability, performed probabilistic decoding could tolerate false detection. analyzed performance through full-factorial simulation. SVM-based touch together with 2D elastic (elasticity = 2) achieve optimal top-5 accuracy 94.2%. evaluation user study, reached speed 26.1 WPM 3.2% uncorrected word-level error rate, significantly better than tap-based gesture-based baseline techniques. Also, proposed technique received highest preference score users, proving its usability real tasks.