One-Touch Pose Detection on Touchscreen Smartphones
Karsten Seipp, Kate Devlin · 2015
We present a technique that allows distinguishing between index finger and thumb input on touchscreen phones, achieving an average accuracy of 82.6% in a real-life application with only a single touch. We divide the screen into a virtual grid of 9mmx9mm units and use a dedicated set of training data and algorithms for classifying new touches in each screen location. Further, we present correlations between physical and digital touch properties to extend previous work.