Modeling the Uncertainty in Pointing of Moving Targets with Arbitrary Shapes

When we try to acquire moving targets such as shooting enemies in computer games, the shapes of these targets are often varied. Considering the effects of target shape in moving target selection is essential for predicting user performances such as error rate in user interfaces involving dynamic content. In this paper, we propose a model to be descriptive of the endpoint uncertainty in pointing of moving targets with arbitrary shapes. The model combines the Gaussian mixture model (GMM) with a Ternary-Gaussian model to describe the impacts of target shape and target motion on selection endpoints of moving targets. Compared to the-state-of-the-art, our model achieved higher performance in the fitting of endpoint distribution and predicting selection error rate.