Design and Evaluation of EdgeWrite Alphabets for Round Face Smartwatches
Keiichi Ueno, Kentaro Go, Yuichiro Kinoshita · 2016
This study presents a project aimed at designing and evaluating a unistroke gesture set of alphanumeric characters targeting round-face smartwatches. We conducted a user study with 10 participants to generate the basic gesture design for 40 characters. For each character, we measured the preference and agreement scores and uncovered any challenges faced in designing unistroke gestures for round-face smartwatches. We developed a gesture recognizer using machine learning, which used a backpropagation mechanism to evaluate the designed gestures. Using the gesture recognizer, we collected 80,000 gesture data, and evaluated them with 5-fold cross-validation. The obtained mean recognition rate was 92.14%.