A user can provide input to a computing device by moving a feature or object, such as a user's finger, within a field of view of at least one imaging element of the computing device. In order to ensure an accuracy of the determined input, the computing device can also attempt to determine a point of view of the user, such as by determining a relative position of the user's face or eyes. By determining a three-dimensional position of a feature and the user's point of view, a three-dimensional vector or other directional information can be determined whereby the intersection of that vector with the computing device indicates an intended location of input corresponding to the feature from the user's point of view.