Improving Touchscreen Typing Using Back-of-Device Grip Interactions

Typing on touchscreen keyboards is inherently inaccurate as users tend to touch locations offset from their intended target. Offsets are user-specific and can further differ for a given user between postures (left-hand, right-hand, indexfinger or two-thumb typing). On the other hand, back-ofdevice interaction has been used to predict screen touches on randomised abstract targets. We propose a new approach where unique offset models are learned for each posture and back-of-device is used to predict posture. Offset models are learned using linear regression while classification is achieved through SVMs and GPs. The models lead to significant improvements in typing accuracy. We show that modelling thumbs within two-thumb typing as separate postures can further improve accuracy. Posture misclassifications, however, can introduce errors in the model. This can be alleviated by weighted averaging of offsets based on probabilities from a probabilistic classifier GP.