I See Your Point
Florian Roider, Tom Gross · 2018
Mid-air pointing gestures enable drivers to interact with a wide range of vehicle functions, without requiring drivers to learn a specific set of gestures. A sufficient pointing accuracy is needed, so that targeted elements can be correctly identified. However, people make relatively large pointing errors, especially in demanding situations such as driving a car. Eye-gaze provides additional information about the drivers' focus of attention that can be used to compensate imprecise pointing. We present a practical implementation of an algorithm that integrates gaze data, in order to increase the accuracy of pointing gestures. A user experiment with 91 participants showed that our approach led to an overall increase of pointing accuracy. However, the benefits depended on the participants' initial gesture performance and on the position of the target elements. The results indicate a great potential to support gesture accuracy, but also the need for a more sophisticated fusion algorithm.